A Governing Water Initiative · Est. 2026 · Published Annually
The Forecast Series
Calibrated readings of trajectories already in motion
In 2026, Governing Water launched an annual forecast series examining the forces reshaping food, water, governance, and the north. Each report is published once a year, written by Palash Sanyal, and grounded in the intersection of systems thinking, governance practice, and frontline advisory work across Canada and internationally.
The forecasts are calibrated readings of trajectories already in motion, published to help leaders, boards, and organizations act before the pressure arrives. They are not predictions of certainty. They are not legal advice, financial advice, or professional governance counsel. They are about understanding what is already in motion.
Each report covers a different domain. Together they form a systems-level view of the decade ahead, written from the perspective of someone who has spent years working at the intersection of water security, governance, Indigenous engagement, and international development.
How We Forecast
Each prediction is grounded in a trend already underway. The forecast extends that trend to its logical consequence within the decade, accounting for the forces most likely to accelerate or complicate it.
The most useful predictions are the ones that create discomfort. If a forecast does not challenge an assumption someone is currently making, it is probably not saying anything worth saying.
Every forecast in the series includes the strongest objection to the prediction and the reasoning behind holding the position anyway. Stress-testing the forecast is part of the forecast.
Numbers and dates in these reports are authorial projections. Where data underpins a figure, the source is named. Where a number is a forecast, it is labelled as one. The distinction matters.
© 2026 Palash Sanyal, CEO, Governing Water Consulting. All rights reserved. These are forecasts prepared by Palash Sanyal, Governing Water Consulting. They are not legal advice, financial advice, investment advice, or professional governance counsel. They are not judgments about whether organizations or individuals are good or bad. They are about understanding what is already in motion. No part of this series may be reproduced, distributed, or transmitted in any form without prior written permission from Governing Water Consulting.
Governing Water · Forecast Series · 2026
The Future
of Food
Thirteen bold predictions reshaping what we eat, how we grow it, and who decides.
Food sits at the intersection of water, climate, technology, power, and culture. Every force reshaping the world reshapes what ends up on the plate. These forecasts follow those forces where they lead, including the places most food conversations stop short of.
Precision Fermentation Displaces the Conventional Farm
Microbes engineered to produce animal proteins, dairy, egg whites, and collagen will reach mainstream grocery shelves at price parity with conventional counterparts. The shift will not announce itself dramatically. Precision-fermented ingredients will become invisible workhorses in everyday foods, indistinguishable from animal-sourced alternatives but produced without a single animal. The transition will be economic before it is ideological, driven by cost curves that conventional agriculture cannot match at scale.
Your DNA Writes Your Menu
Nutrigenomic platforms will move from novelty to necessity. Apps combining real-time blood glucose monitoring, microbiome sequencing, and genetic profiles will generate hyper-personalized meal plans with clinical precision. Restaurants will begin offering biosync menus that adapt to individual diner profiles. The implications extend well beyond wellness: food becomes a medical intervention, and the line between a restaurant meal and a prescription blurs in ways that regulators are nowhere near ready to address.
Every flood, every drought, every empty shelf is the food system speaking in the only language it has. Water. We have been receiving the signal for decades. The forecasts below are about what happens when we finally read it.
Palash Sanyal · CEO, Governing Water Consulting
Weight Loss Drugs Will Reshape the Entire Food Industry Before Anyone Plans For It
GLP-1 drugs including Ozempic and Wegovy reduce appetite at a scale and speed that the food industry has never encountered. People on these medications eat less, crave differently, and abandon categories entirely, particularly ultra-processed snacks, sugary beverages, and fast food. The implications for food manufacturers, grocery retailers, and restaurant chains are structural and already underway. Companies that built their revenue models on volume consumption are watching their core customer behaviour change in real time. Caloric intake is falling among a growing segment of the population and the food system was built to feed people who wanted more, not less.
The Sober Curious Movement Becomes a Culinary Category
Functional, alcohol-free beverages powered by adaptogens, nootropics, and nano-cannabinoids will claim a significant share of the bar experience. Premium mindful drinking will evolve into its own culinary category, with dedicated zero-proof sommeliers and tasting menus at fine dining restaurants. The movement is generational, economic, and health-driven simultaneously. Venues that treat non-alcoholic options as an afterthought will lose a growing segment of high-spending diners who have simply stopped apologizing for not drinking.
Regenerative Agriculture Earns a Luxury Premium
Soil-positive, carbon-sequestering farming practices will become a verified luxury differentiator. Consumers will pay a meaningful premium for traceable regenerative labels backed by blockchain verification, following the same trajectory that organic certification took a generation ago. Legacy brands that fail to transition will face structural demand decline among younger consumers for whom soil health and climate accountability are purchasing criteria, not marketing noise. The premium will shrink as the practice scales, but the first movers will have built durable brand equity that late adopters cannot buy.
AI Manages Industrial Kitchens Before Chefs Realize the Shift Has Happened
Artificial intelligence will overhaul food production and restaurant operations, optimizing recipes to minimize waste, predicting demand to eliminate spoilage, and customizing flavour profiles at scale. The change will arrive incrementally through procurement software, inventory systems, and recipe optimization tools before anyone frames it as a kitchen revolution. The culinary workforce implications will be significant and largely unaddressed by current labour policy. The phrase chef-driven will acquire a new meaning that most current chefs did not agree to.
Hyper-Regional Cuisines Drive the Next Wave of Fine Dining
In reaction to globalized palates and homogenized menus, ultra-specific micro-regional cuisines will command outsized cultural cachet. Chefs who master depth over breadth, anchored to a single valley, coastline, or community food tradition, will define the next era of serious dining. The movement is already visible in Nordic and Peruvian fine dining. It will spread to Indigenous Canadian, West African, and Central Asian culinary traditions as the gatekeeping infrastructure of the global restaurant industry continues to decentralize.
Plastic-Free Packaging Becomes a Legal Requirement
Regulatory pressure across the EU, UK, and eventually North America will mandate biodegradable, compostable, or reusable packaging across the entire food supply chain. Brands that build circular packaging ecosystems now will gain lasting competitive advantages as compliance costs crush slower competitors. The transition will be uneven, with large corporations absorbing costs that will devastate smaller suppliers who lack the capital to retool. The environmental benefit will be real. The consolidation effect on the food industry will be significant and largely unremarked.
Gut Health Becomes the Defining Wellness Claim of the Decade
As microbiome science matures, gut-positive will become the central wellness claim across every food and beverage category, eclipsing organic and non-GMO in consumer priority. Fermented foods, prebiotic fibres, and postbiotic ingredients will anchor new product launches at every price point. The science is real but the marketing will significantly outpace the evidence, creating a regulatory challenge that food safety bodies are not resourced to address at the speed the market is moving.
Vertical Farms Become Standard Urban Infrastructure
High-density, AI-managed vertical farms will become embedded in grocery stores, hospitals, hotels, and airports as standard urban infrastructure. Year-round hyper-local produce will break the seasonal constraint that has defined agriculture since the beginning of settled human civilization, delivering fresher food with a fraction of conventional water use. The technology is proven. The scaling challenge is economic and political, requiring land use policy changes that urban governments have been slow to make and investors have been faster to push for.
Dining Becomes Immersive Theater
The boundary between restaurant and entertainment venue will dissolve. Multisensory dining employing scent, sound, projection, and haptics will become a recognized art form, with the most coveted tables booked like theatre seats. Food will be the medium. Experience will be the product. The economics will reward a small number of venues spectacularly while concentrating fine dining further into cities that can sustain the capital requirements. Most of this innovation will happen in places already well-served by restaurants, which raises questions about who the food system is actually innovating for.
Food Insecurity Will Be Rebranded as a Logistics Problem to Avoid Calling It a Justice Problem
As food technology investment accelerates in wealthy urban markets, the gap between food innovation and food access will widen. Governments and corporations will increasingly frame food insecurity as a distribution and supply chain challenge, funding technology solutions that serve markets already well-supplied while communities with chronic access gaps remain outside the investment thesis. The language of food systems innovation will be appropriated by actors whose products reach the people who need them least. This is already happening. The forecast is that it will be named less honestly as it scales.
Water Scarcity Will Determine What Gets Grown and Where Before Any Policy Does
Agricultural water stress will restructure global food production geography faster than any trade agreement, subsidy program, or food policy framework. Crops will migrate. Farming regions will collapse. New growing zones will emerge in areas with no agricultural infrastructure or workforce. The decisions will be made by water availability, not planning. Governments will respond after the shift has already happened, designing policy for a food system that moved without them. The communities most affected will be the ones with the least capacity to adapt and the least voice in the decisions that shaped the outcome.
© 2026 Palash Sanyal, CEO, Governing Water Consulting. All rights reserved. These are forecasts prepared by Palash Sanyal, Governing Water Consulting. They are not legal advice, financial advice, investment advice, or professional governance counsel. They are not judgments about whether organizations or individuals are good or bad. They are about understanding what is already in motion. No part of this report may be reproduced, distributed, or transmitted in any form without prior written permission from Governing Water Consulting.
Governing Water · Forecast Series · 2026
The Future
of Water
Thirteen bold predictions reshaping how the world governs, produces, and survives on water.
Water is not backdrop. In the decade ahead, it becomes the central variable in every calculation: food, energy, migration, sovereignty, and conflict. These are not distant possibilities. They are already underway.
Floating Food Platforms Replace the Vulnerable Field
As climate volatility renders ground-level agriculture increasingly unpredictable, a new generation of offshore and inland floating platforms will emerge as climate-resilient food production zones. Think cruise-ship scale agri-infrastructure: controlled environments, decoupled from soil and rainfall, producing staple crops year-round. Nations with coastlines and freshwater lakes will pilot these first for high-value specialty crops, then at scale for food security. Within this decade, growing food in the ground will begin to be seen as the riskier option.
Canada Will Sell Water to the United States
The idea has been politically toxic for decades. It will not stay that way. As the American West enters permanent drought and Midwest aquifer depletion accelerates, pressure on Canada to commodify its freshwater surplus will become a defining bilateral issue. A federal government facing trade leverage and economic incentives will quietly open the door, beginning with small-scale agreements framed as “water sharing” rather than export. The legal and constitutional barriers are real but they have been eroded before when the economics are large enough.
Every flood, every drought, every dry well is climate change speaking in the only language it has. Water. We have not been listening.
Palash Sanyal · CEO, Governing Water Consulting
Bottled Water Companies Face Nationalization in Three Countries
The political backlash against private extraction of public water is already building. By 2032, at least one middle-income country facing acute urban water stress will nationalize private water bottling and distribution operations, citing the public trust doctrine. Others will follow. Global brands with extraction licenses in water-scarce regions will face a legal environment that looks nothing like the one they signed contracts in. Investors are not pricing this risk anywhere close to accurately.
The First Declared Water War of the 21st Century
Not a conflict in which water plays a role. An actual armed conflict in which water access is the stated, primary cause. The candidates are well known: Ethiopia and Egypt over the Grand Ethiopian Renaissance Dam, India and Pakistan over Indus tributaries, or Central Asian states over Amu Darya allocation. Analysts have warned about this for thirty years. The window in which diplomacy alone can hold it off is closing. The decade from 2026 to 2035 is the most dangerous period in recorded history for water-related interstate conflict.
AI-Driven Irrigation Cuts Agricultural Water Use by 40%
Precision irrigation managed by machine learning, integrating soil sensors, weather prediction, satellite imagery, and crop physiology models, will move from pilot projects to standard practice across large-scale irrigated agriculture. As producers across Saskatchewan already understand, the economics of irrigation only work when efficiency is maximized. AI will close the last mile of optimization that human scheduling never could, reducing consumptive water use dramatically while lifting yields. The transition will be faster and more disruptive to existing agronomy service industries than anyone is currently planning for.
Cities Will Ration Water by Income and Call It Policy
As municipal water systems in mid-sized cities across the Global South buckle under demand, tiered pricing will evolve into something far less palatable: volumetric rationing tied to economic status. Wealthy districts will maintain pressure and availability. Poor districts will face scheduled outages rationalized as conservation measures. Already happening informally in Cape Town, Chennai, and Mexico City, by 2030 this pattern will be codified in city water bylaws and defended in court as necessary triage. The human rights implications will be staggering and largely ignored.
Meltwater Governance Becomes the Arctic’s Most Urgent Frontier
As glaciers and permafrost release centuries of stored freshwater at unprecedented rates, the legal and governance frameworks to manage that flow simply do not exist. Nunavut, the Northwest Territories, and Alaska will face competing claims from downstream communities, industry, and ecosystems. New regulatory architecture, built with Inuit and Arctic Indigenous communities at the centre, will need to be designed from scratch. The window to do this collaboratively, before crisis forces a different kind of decision-making, is narrow and closing fast.
Weather Modification Programs Trigger International Disputes
Cloud seeding is already practiced by over fifty countries. In the next decade, as drought intensifies, nations will scale these programs dramatically, and the downstream effects on neighbouring countries will generate legal and diplomatic crises. When China seeds clouds over the Tibetan Plateau and India receives less monsoon rainfall, or when Gulf states seed over the Arabian Sea and divert moisture from the Horn of Africa, the question of who owns precipitation will move from academic to urgent. There is no international legal framework for this. Building one will take longer than the crises it needs to address.
Water Disclosure Becomes Mandatory for Public Companies
Following the trajectory of carbon disclosure, regulatory bodies in the EU, UK, and Canada will mandate that publicly traded companies report water risk, consumption, and watershed impact in annual filings. Institutional investors will integrate water stewardship into ESG scoring with the same rigour now applied to emissions. Companies operating in water-stressed basins without credible transition plans will face rising cost of capital and eventual divestment pressure comparable to fossil fuel exposure today. The companies most exposed are not the ones most people would guess.
The Gulf States Will Abandon Domestic Agriculture to Survive
Saudi Arabia, UAE, Kuwait, and Qatar have been quietly exiting domestic food production for years, having nearly exhausted their fossil aquifers on wheat and alfalfa. The next phase is more radical: full abandonment of in-country agriculture in favour of controlling farmland and water rights in Africa, Central Asia, and South America. These nations are not food insecure yet. They are choosing to become agricultural colonizers before they become agricultural casualties. The communities displaced by those land and water acquisitions will not have a vote in any of it.
© 2026 Palash Sanyal, CEO, Governing Water Consulting. All rights reserved. These are forecasts prepared by Palash Sanyal, Governing Water Consulting. They are not legal advice, financial advice, investment advice, or professional governance counsel. They are not judgments about whether organizations or individuals are good or bad. They are about understanding what is already in motion. No part of this report may be reproduced, distributed, or transmitted in any form without prior written permission from Governing Water Consulting.
Governing Water · Forecast Series · 2026
The Future
of Governance
Thirteen bold predictions reshaping who holds power, how decisions get made, and what legitimacy will mean.
Governance sits at the intersection of power and trust. That intersection is shifting in ways that formal institutions are struggling to track. What emerges over the next decade will look unfamiliar to anyone trained on the frameworks of the last one.
Boards Will Be Held Personally Liable for Decisions They Cannot Explain
The era of boards hiding behind process will end. Courts, regulators, and the public will increasingly demand that directors demonstrate they genuinely understood the decision they made and its consequences. The shift toward generative accountability, already visible in climate litigation and social licence disputes, will accelerate. Directors who cannot articulate the reasoning behind major decisions will face personal exposure alongside institutional exposure. Following a procedure will no longer constitute sufficient governance.
Indigenous Governance Models Will Become the Template, Not the Exception
Western governance structures were built on assumptions of speed, hierarchy, and extraction. Those assumptions are failing in real time. Indigenous governance frameworks, rooted in long-term stewardship, consensus, and relational accountability, will be adopted not as a cultural accommodation but as a practical necessity. Organizations navigating land, water, climate, and community trust will find that the models they need already exist. They simply were not the ones writing governance textbooks.
Most boards manage compliance. The harder and rarer skill is choosing under genuine uncertainty, when the process has run out and a decision still has to be made.
Palash Sanyal · CEO, Governing Water Consulting
Informal Power Will Permanently Outpace Formal Authority
The most consequential governance decisions of the next decade will not happen in boardrooms or legislatures. They will happen in private networks, messaging groups, and behind closed agreements that leave no public record. Formal institutions will increasingly ratify decisions already made elsewhere. The pattern is well documented across international finance, corporate deal-making, and political negotiation. What operated as an exception in some sectors will become the standard operating mode across corporate, nonprofit, and public governance. The question is whether organizations design for this honestly or pretend the formal process is still where power lives.
Psychological Safety Will Become a Governance Requirement, Not a HR Initiative
Boards and leadership teams that cannot have honest internal disagreement will make structurally worse decisions. The research is clear: where people cannot speak safely, critical information does not reach decision-makers in time. Regulators and governance codes will begin requiring boards to demonstrate that dissent is structurally enabled, visible in how meetings are run and how minority views are recorded. Organizations treating psychological safety as a leadership program will be caught unprepared when it becomes a governance audit item.
The Nonprofit Sector Will Face a Legitimacy Crisis It Has No Framework to Survive
Nonprofits have operated under a governance exemption for decades: mission replaces accountability, good intentions substitute for performance, and boards of well-meaning volunteers are treated as sufficient oversight. That exemption is running out. Donors, funders, and communities are demanding rigour from nonprofits comparable to what they expect from public institutions. Organizations that cannot demonstrate genuine governance will lose funding, trust, and relevance faster than they can adapt. Most nonprofit boards have never had that conversation honestly.
AI Will Sit at the Governance Table Before Anyone Has Agreed It Should
Boards and senior leadership teams are already using AI to synthesize board packages, model risk scenarios, and draft governance documents. Most have not had a single conversation about what that means for accountability, judgment, and the legal definition of a decision. Within this decade, the question of whether an AI-assisted decision constitutes a valid board resolution will reach courts and regulators. Organizations that wait for regulatory clarity before developing their own AI governance policies will find the policy written for them under circumstances they did not choose.
The Next Generation of Directors Will Refuse the Governance They Inherited
Younger leaders entering board roles carry fundamentally different expectations about transparency, purpose, and accountability. They did not build the governance structures they are joining. The generational collision on boards will go deeper than communication styles or technology comfort. It will surface the question of whether inherited governance models serve the organizations they are supposed to protect or primarily protect the people who built them. Boards that cannot answer that question honestly will lose their next generation of directors before those directors reach the table.
Reputation Will Replace Regulation as the Primary Governance Enforcement Mechanism
Formal regulatory enforcement is slow, expensive, and inconsistent. Social media has created a parallel enforcement system that is fast, unforgiving, and operates without due process. Boards and executives who once had years to respond to governance failures now have hours. Cancel culture, brand boycotts, and viral accountability are not threats that exist outside governance. They are governance, operating at a speed and scale that formal institutions cannot match. Organizations that treat reputational risk as a communications problem rather than a governance problem will learn the difference at the worst possible moment.
University Governance Will Collapse Under Its Own Weight
Academic governance was designed for a world that no longer exists. Shared governance models built on faculty senates, committee structures, and deliberate consensus are structurally incapable of responding at the speed that financial, demographic, and technological pressures now demand. Universities that cannot restructure their governance will not survive the next decade intact. Mergers, closures, and forced transformations will accelerate, and the governance failures that precede them will follow a recognizable pattern: too many voices, too little authority, and no one willing to make the decision that needed to be made three years earlier.
Governance Will Become a Product Before It Becomes a Standard
The governance advisory industry is fragmented, credential-heavy, and largely inaccessible to the organizations that need it most. Small nonprofits, Indigenous organizations, municipalities, and community boards cannot afford the consultants designing governance for large corporations. The response will come from entrepreneurs who productize governance tools, decision frameworks, and board support systems at a fraction of the traditional cost. The next generation of governance innovation will look less like a professional services firm and more like a platform. Organizations waiting for a credentialed consultant will be outpaced by the ones willing to use a well-designed tool.
Public Trust Will Become a Balance Sheet Item
Organizations have long treated public trust as a communications outcome: something you manage after a crisis, not something you measure before one. That is changing. Insurers, institutional funders, and credit rating agencies will begin requiring organizations to demonstrate measurable trust indicators as a condition of coverage, investment, and access to capital. The organizations that cannot produce that evidence will carry a visible risk premium. Trust will stop being a value and start being a number, with all the discipline and accountability that implies.
The Board Table Will Become Where Unresolvable Social Tensions Land First
Boards were designed to govern organizations. They were not designed to adjudicate polarized communities, navigate identity conflicts, or hold the weight of social movements that governments refuse to address. That distinction is collapsing. As public institutions retreat from contested social ground, the pressure transfers to the organizations closest to communities: nonprofits, Indigenous bodies, universities, municipal agencies. Their boards will face decisions for which no governance framework, no policy manual, and no training program has prepared them. Most will handle it badly. The ones that do not will have done the work to understand the difference between a governance question and a social one.
Governance Credentials Will Lose Value Faster Than the Bodies Issuing Them Can Adapt
The governance credential industry is built on the premise that certification signals competence. That premise is weakening. Organizations experiencing real governance failures are full of credentialed directors. Organizations making exceptional decisions often have people around the table who hold no formal designation at all. As this pattern becomes harder to ignore, the market will shift toward demonstrated judgment: track records, peer reputation, and observable decision-making under pressure. Credential bodies will respond with new frameworks and updated curricula. They will be responding to a world that has already moved.
The Pushbacks and the Responses
Every forecast deserves to be challenged. These are the strongest objections to each prediction, and the reasoning behind holding the position anyway.
Personal Liability for Unexplainable Decisions
The business judgment rule exists precisely to protect directors from liability for good-faith decisions that turned out badly. You cannot hold someone liable for the quality of their reasoning, only the process they followed.
The business judgment rule was designed for commercial risk. It was not designed for governance failure under sustained public and regulatory scrutiny. The standard is shifting, and courts in multiple jurisdictions are already narrowing the protection it once offered.
Indigenous Governance as the Template
Lifting governance models out of their relational and land-based context strips them of meaning. Indigenous consensus frameworks work in communities with shared history. Boards of strangers meeting quarterly is a fundamentally different context.
The prediction is about adopting underlying principles: long-term thinking, stewardship over extraction, accountability to those affected rather than those invested. Those principles translate across contexts even when the form does not.
Informal Power Outpacing Formal Authority
This has always been true. Publishing it as a forecast risks legitimizing shadow governance rather than reforming it. Some readers will ask whether this is a description or an endorsement.
The scale and speed are genuinely new, even if the pattern is not. And the point is exactly the opposite of endorsement: organizations that pretend formal process is where power lives will keep being surprised. Naming the pattern is the first step toward designing accountability into it.
Psychological Safety as a Governance Requirement
Psychological safety is a team culture metric. Regulators have no business auditing how safe people feel in a boardroom. This conflates leadership culture with fiduciary duty.
When board culture systematically suppresses dissent and that suppression contributes to a governance failure with material consequences, the distinction between culture and fiduciary duty collapses. Regulators will follow the liability, not the category.
Nonprofit Legitimacy Crisis
Underfunding, not under-governance, is the real crisis. This prediction blames organizations for systemic failures caused by inadequate public investment. That is unfair to people doing real work with very little.
Both are true simultaneously. Underfunding is real and structural. Governance failure is also real and within organizations’ control. Funders will increasingly use governance as the criterion for scarce dollars, which makes governance a survival issue regardless of the underlying inequity.
AI at the Governance Table
Existing fiduciary duty frameworks already handle this. Directors are responsible for whatever inputs they use. AI is a tool, and tools do not change the duty of care.
The frameworks were written before a tool could synthesize a 400-page board package into a two-page summary that a director reads without knowing what was removed. The duty may be unchanged. The conditions under which it is exercised have changed completely.
Next Generation Directors Refusing Inherited Governance
Generational generalizations are reductive. Many younger directors are deeply conservative about governance and many of the most radical reformers are in their sixties.
Fair. The prediction is less about age and more about the values cohort entering governance now. The stronger version of the argument is about purpose alignment and tolerance for opacity, both of which are shifting in ways that correlate with but are not determined by generation.
Reputation Replacing Regulation
Reputational enforcement without due process operates without recourse for those targeted wrongly. That is a genuine harm, and any honest forecast has to name it.
The pushback is correct and the prediction does not resolve that tension. The forecast is descriptive, not prescriptive. The point is that boards need to govern as if this is already true, because it already is. Whether it should be true is a separate and important question.
University Governance Collapse
This prediction has been made every decade for fifty years and universities keep surviving. Many govern badly and remain financially stable. Governance dysfunction and institutional collapse follow different timelines.
They have survived in part because protected public funding masked governance dysfunction. That protection is eroding faster than governance is reforming. The prediction is about the convergence of multiple pressures at once, not governance failure in isolation.
Governance as a Product
Governance judgment cannot be productized without losing what makes it valuable. A template replaces neither advice nor a sustained relationship. The risk is that access gets confused with quality.
Most organizations never had access to quality governance advice to begin with. A well-designed tool used consistently is better than no governance support at all, which is the current reality for the majority of small and community organizations. The ceiling matters less than the floor.
Public Trust as a Balance Sheet Item
Trust is not measurable in any meaningful way. Any metric you create will be gamed. Organizations will manufacture trust scores the same way they manufacture sustainability ratings.
Every important governance measure gets gamed eventually. That does not make measurement worthless. It makes rigorous methodology essential. The alternative, treating trust as immeasurable and therefore unaccountable, has already produced the environment we are in.
The Board Table as Landing Zone for Social Tensions
Boards that take on social conflict beyond their mandate are overstepping. Their job is to govern the organization, not adjudicate society. Mission creep is itself a governance failure.
The choice is not between engaging and not engaging. When a social conflict arrives at an organization’s door, the board is already involved whether it has chosen to be or not. The governance question is how to respond with clarity about what the board can and cannot decide, not whether to respond at all.
Governance Credentials Losing Value
Credentials exist to create a baseline. Without them, board recruitment becomes a network game where the well-connected get seats regardless of competence. Credentials democratize access more than they restrict it.
That argument holds when credentials reliably signal competence. The prediction is precisely that the signal is weakening. If credentials no longer reliably predict governance quality, then the democratization argument dissolves alongside the competence one. What replaces them needs to be designed intentionally, which is the harder and more important work.
© 2026 Palash Sanyal, CEO, Governing Water Consulting. All rights reserved. These are forecasts prepared by Palash Sanyal, Governing Water Consulting. They are not legal advice, financial advice, investment advice, or professional governance counsel. They are not judgments about whether organizations or individuals are good or bad. They are about understanding what is already in motion. No part of this report may be reproduced, distributed, or transmitted in any form without prior written permission from Governing Water Consulting.
Governing Water · Forecast Series · 2026
The Future
of Work
Eleven bold predictions reshaping who works, why they work, and what work does to them.
Every major consulting firm has published its version of this forecast. They cover AI, reskilling, and hybrid models. What they do not cover is harder to look at: meaning, collapse, liability, and the slow unravelling of the contract between employers and the people they depend on. These predictions go there.
The Meaning Crisis at Work Will Become the Defining Management Problem of the Decade
Productivity tools, AI assistants, and reskilling programs address what people do at work. None of them address why it should matter. The workforce entering organizations now carries a deeper question than any previous generation has presented to employers at scale: what is this for? Engagement metrics have been declining for a decade and no intervention has reversed the trend. The reason is that organizations are solving for motivation while the actual problem is meaning. These are not the same thing. The companies that figure out the difference will retain talent and generate discretionary effort that no incentive structure can manufacture. The ones that do not will spend the next decade confused about why their people keep leaving for less money.
Men Will Become the Overlooked Workforce Crisis Nobody Is Willing to Name
Labour force participation among men without post-secondary credentials has been declining for decades. The crisis is not economic, it is structural and psychological. Work provided identity, purpose, community, and status for generations of men who had no other institutional framework for those things. As those jobs disappeared and were not replaced by new ones that carry comparable meaning, a generation of men has drifted into economic marginality without a narrative for why or a pathway back. The workforce conversation is almost entirely oriented around inclusion and representation for groups that are underrepresented in high-status work. That is necessary. What is also necessary is an honest conversation about the men who have already left, the boys who are not showing up, and the downstream consequences for families, communities, and economies of a permanent male underclass. The silence on this is ideological, and it is costing more than anyone is admitting.
The reskilling conversation assumes people want to be reskilled. What it does not ask is whether the work waiting at the end of the reskilling program is worth showing up for.
Palash Sanyal · CEO, Governing Water Consulting
The Fifty-Year Career Will Destroy Every Retirement Framework Built to Replace It
People are not retiring at 65. They are not retiring at 70. A significant share are not retiring in any recognizable sense at all. The combination of longer healthspans, inadequate pension accumulation, and the psychological devastation of sudden disengagement from purposeful work means that the clean line between career and retirement is dissolving. What is emerging in its place has no policy framework, no HR architecture, and no cultural script. Organizations are built for a workforce that arrives young, rises steadily, and exits cleanly. The workforce that is actually arriving, which includes people in their seventies contributing at senior levels alongside people in their twenties who expect portfolio careers from day one, fits none of those assumptions. The companies that redesign for the fifty-year career will have access to talent their competitors cannot hold.
Employers Will Be Held Legally Liable for Algorithmic Harm to Workers Before They Have Policies to Prevent It
AI systems are already making decisions that affect workers: who gets interviewed, who gets scheduled, who gets flagged for performance review, whose contract is renewed. Most of these systems operate without meaningful oversight, without disclosure to workers, and without any accountability framework. Courts and regulators in the EU and Canada are moving faster than organizations are. The first major rulings against employers for algorithmic harm to workers, discriminatory scheduling, biased performance scoring, illegal dismissal triggered by automated systems, will arrive before most organizations have even audited what their systems are doing. The companies that have done the governance work will be fine. The ones that have not will be making very expensive examples for everyone else.
Workplace Loneliness Will Be Recognized as an Occupational Health Crisis, Not a Personal Problem
Remote and hybrid work solved a commuting problem and created a connection problem that nobody has honestly priced. The social infrastructure of the office, incidental conversation, observed effort, informal mentorship, the ambient knowledge that you are not alone, cannot be replicated on a video call, and most organizations have not tried. The evidence linking chronic workplace loneliness to depression, cardiovascular disease, cognitive decline, and early mortality is not new. What is new is the scale. A generation of workers has now spent formative career years in social isolation, and the downstream effects on mental health, professional development, and organizational loyalty are only beginning to register. Employers who treat this as a personal wellness issue will be caught off guard when it becomes a workers’ compensation and duty-of-care question.
The Gig Economy Will Split Into Two Completely Different Things That Share a Name
Independent work is not one thing. At the top, a growing class of high-earning professionals, architects, consultants, engineers, educators, researchers, are building portfolio careers that deliver more autonomy, higher income, and greater meaning than any employment arrangement they were offered. At the bottom, a growing mass of platform-dependent workers, delivery drivers, care workers, micro-task completers, are trapped in structures that extract maximum labour at minimum cost with no floor on what they can lose. Both groups are called gig workers. They have almost nothing in common. The policy conversation treats them as one population and thereby serves neither. The divergence will accelerate. The people at the top will be the most economically secure independent professionals in history. The people at the bottom will be the most exploited workforce without legal status since the industrial revolution.
The Psychological Contract Between Employer and Employee Will Not Be Repaired. Something Else Will Replace It.
The old contract offered security in exchange for loyalty. That contract broke during the downsizing era of the 1990s and has not recovered. The new generation entering the workforce did not experience the original contract and does not mourn its absence. What they are negotiating instead is something more transactional, more explicit, and in many ways more honest: defined terms, visible expectations, and no pretense that the organization cares about them beyond what they deliver. This is neither good nor bad. It is real. The organizations still trying to rebuild loyalty through culture, ping-pong tables, and mission statements are misreading the moment. The ones that are thriving are those that have designed explicit, honest, and fair transactional relationships and stopped pretending to be families.
Organizations are not families. The workers who figured that out first are the ones who are thriving. The ones still waiting to be appreciated are the ones burning out.
Palash Sanyal · CEO, Governing Water Consulting
Middle Management Will Not Survive the Current Decade in Its Current Form
AI does not just automate tasks. It automates coordination, synthesis, status reporting, meeting facilitation, and information relay. Which is to say, it automates most of what middle management actually does. The layer of organizational hierarchy whose function is to translate strategy downward and report performance upward is the most directly threatened by AI of any workforce segment, and it is the one that has received the least attention in the reskilling conversation. The roles will not disappear overnight. What will happen first is that the number of people required to perform those functions will shrink dramatically, and the skills required of those who remain will be almost entirely about judgment and relationship rather than process and oversight. Most current middle managers have been trained for the version that is leaving.
The Global South Will Build the Workforce Infrastructure the North Dismantled
Africa’s working-age population will double by 2050. India is adding labour force entrants at a pace no other economy can match. These are not abstract demographic statistics. They are the raw material of the most consequential workforce story of the century. While Canada, Germany, Japan, and the United States manage the politics of labour scarcity, aging, and automation, the countries that build strong educational infrastructure, portable credentialing systems, and digital-first labour markets will capture a disproportionate share of the global knowledge economy. The north has been offshoring physical labour for decades. What is coming next is the offshoring of judgment work, legal analysis, financial modelling, engineering design, medical diagnosis, at a scale that will restructure professional labour markets in ways that no union, no immigration policy, and no trade agreement was designed to handle.
The University Degree Will Lose Its Employment Signal Before Any Alternative Is Ready to Replace It
Employers are already hiring on demonstrated skill, portfolio work, and micro-credentials at rates that would have been unthinkable a decade ago. The university degree is losing its function as a hiring filter, and the decline will accelerate as AI tools make it trivially easy to produce degree-equivalent outputs without one. What is not ready is the replacement. Skills-based hiring sounds straightforward in a consulting presentation. In practice, assessing judgment, learning capacity, and cultural fit without the proxy of a recognized credential is expensive, inconsistent, and biased in ways that credentials were at least designed to address. The transition period, in which the old signal no longer works and the new one does not yet exist, will disadvantage exactly the workers the skills-based movement claims to liberate.
Indigenous Economic Development Will Become the Most Significant Workforce Story in the Canadian North
Resource development, infrastructure investment, and land rights settlements are creating Indigenous-led economic institutions at a scale and pace that has no precedent in Canadian history. The workforce implications are profound and almost entirely unaddressed in national labour policy. First Nations, Métis, and Inuit communities are building institutions that will employ tens of thousands of people in roles that require governance expertise, financial management, technical operations, and community leadership. The training systems, credentialing pathways, and workforce development infrastructure to support this do not currently exist at the required scale. This is simultaneously one of the most urgent workforce development challenges in Canada and one of the least visible in the mainstream conversation about the future of work.
The Pushbacks and the Responses
Every forecast deserves to be challenged. These are the strongest objections to each prediction, and the reasoning behind holding the position anyway.
The Meaning Crisis as the Defining Management Problem
People have always wanted meaning at work. This is not new. Every generation thinks it has discovered something the previous one missed. What actually drives retention is compensation and job security, which have always mattered more than purpose surveys suggest.
Compensation satisfies the condition for showing up. It does not produce discretionary effort, loyalty, or the kind of sustained contribution that builds organizations. The data on disengagement is a decade old and getting worse despite rising wages. Something else is happening.
Men as the Overlooked Workforce Crisis
Men have historically dominated every economic advantage. Framing male disengagement as a crisis while women, racialized workers, and disabled people still face structural exclusion from the labour market is a misallocation of political and policy attention.
The prediction is not a competition. Both things are simultaneously true. Structural exclusion of historically marginalized groups requires sustained policy attention. So does the demographic collapse of male participation in lower-credential labour markets. Treating them as mutually exclusive is how both go unaddressed.
The Fifty-Year Career Destroying Retirement Frameworks
Most people working past 65 are doing so because they cannot afford to retire, not because they want to. Celebrating the fifty-year career risks normalizing inadequate pension systems and employer extraction of labour from people who have no real choice.
The financial and the volitional are both real and both matter. The prediction is about what organizations need to design for, not what social policy should celebrate. Organizations that assume workers over 60 are winding down will lose the most experienced people in their workforce to competitors who do not make that assumption.
Employer Liability for Algorithmic Harm
Existing employment law already covers discriminatory outcomes regardless of how they are produced. Algorithmic harm is covered under human rights legislation. This does not require new frameworks, just enforcement of existing ones.
Existing frameworks were written before an algorithm could make 50,000 scheduling decisions per week that systematically disadvantage protected groups without any individual decision-maker being aware. The accountability architecture for distributed, opaque, high-volume automated decisions does not exist yet. Courts will be making it up as they go.
Workplace Loneliness as an Occupational Health Crisis
Employers cannot be responsible for workers’ social lives. Remote work also produces real mental health benefits, reduction in commute stress, more time with family, and greater autonomy. The loneliness narrative overweights the costs and ignores the benefits.
The benefits of remote work are real and unevenly distributed. They accrue most to workers with strong home environments, established careers, and existing social networks. The costs are most concentrated among early-career workers, people in precarious housing, and anyone whose primary social infrastructure was the workplace. The average conceals a serious distributional problem.
The Gig Economy Splitting Into Two Different Things
Platform work, even at low wages, provides access to income for people who would otherwise have no formal labour market access. Calling it exploitation ignores that for many workers it is a genuine improvement over the alternative.
Better than no income is a low bar for a labour standard. The prediction is about trajectory: platform work at the bottom is becoming less remunerative, not more, as platform concentration increases and algorithmic wage-setting becomes more sophisticated. Better than nothing today does not mean better than alternatives tomorrow.
The Psychological Contract Will Not Be Repaired
Transactional relationships produce transactional effort. Organizations with genuine culture and genuine loyalty consistently outperform transactional ones. This prediction recommends the wrong response to a real problem.
The distinction is between genuine culture, which is built through consistent behaviour and real accountability, and performed culture, which is a communication strategy. The prediction is against the latter. Organizations with real cultures are not being criticised here. They are the ones that will prove the point.
Middle Management Will Not Survive in Its Current Form
This prediction has been made about every new technology for fifty years and middle management has survived them all by absorbing new functions as fast as old ones disappeared. There is no reason to believe AI is categorically different.
Previous technologies automated physical and routine cognitive tasks. AI automates coordination and synthesis, which is the core function, not a peripheral function, of management. The analogy would be if the previous technologies had automated sales rather than the assembly line. The absorption mechanism does not work when the core function is what is being automated.
The Global South Building Workforce Infrastructure the North Dismantled
Knowledge work offshoring has been predicted before and has consistently run into quality, coordination, and intellectual property challenges that limit its scale. The structural barriers are real and have not been resolved by AI.
AI changes the quality equation fundamentally. The barrier to professional-grade output from a well-educated worker in Lagos or Nairobi was never effort or intelligence, it was access to specialized knowledge, tools, and credentialing systems. AI reduces those barriers across the board. The structural factors that previously limited offshoring of judgment work are being systematically dismantled.
The University Degree Losing Its Employment Signal
Degrees signal more than skill. They signal conscientiousness, completion, the ability to manage sustained effort and ambiguity. Those signals do not become redundant just because the credential can be gamed.
All of that is true, and none of it is what the degree is actually being used for in most hiring processes. Most hiring managers cannot articulate what the degree signals for the specific role. They use it as a filter because everyone else does. When enough employers stop, the filter collapses regardless of the underlying argument for its validity.
Indigenous Economic Development as the Workforce Story of the Canadian North
Indigenous economic development has been promised and underfunded in Canada for generations. The governance and infrastructure gaps are real but so is the political record of failing to close them. Why would this decade be different?
The difference is that this generation of Indigenous economic development is being built from Indigenous-owned institutions, not through government programs. The land rights settlements, resource revenue agreements, and governance structures now in place give communities instruments of economic agency that did not previously exist. The pace of institution-building is already faster than anything that came before. The workforce infrastructure to support it is the gap, and that gap is visible and closeable.
© 2026 Palash Sanyal, CEO, Governing Water Consulting. All rights reserved. These are forecasts prepared by Palash Sanyal, Governing Water Consulting. They are not legal advice, financial advice, investment advice, or professional governance counsel. They are not judgments about whether organizations or individuals are good or bad. They are about understanding what is already in motion. No part of this report may be reproduced, distributed, or transmitted in any form without prior written permission from Governing Water Consulting.
Governing Water · Forecast Series · 2026
The Future
of AI
Twelve bold predictions about what AI will actually do — not to tasks, but to institutions, accountability, and the nature of expertise.
Every consulting firm has published its AI forecast. They cover agents, ROI, reskilling, and regulatory frameworks. What they do not cover is harder: the collapse of expertise as a social category, the liability gap that nobody is pricing, the countries that will win for reasons nobody is talking about, and what happens when AI gets governance wrong at scale. These predictions go there.
AI Will Not Replace Experts. It Will Make Expertise Invisible — Which Is Worse.
The conversation about AI and expertise has been framed as replacement: will AI take the job of the lawyer, the doctor, the analyst? That is the wrong question. What AI is already doing is collapsing the perceived distance between expert and non-expert. When anyone can produce a legal memo, a diagnostic summary, or a financial analysis in seconds, the social and institutional infrastructure built on the scarcity of that output begins to dissolve. The problem is not that the experts disappear. It is that nobody can tell the difference anymore between expert judgment and the confident-sounding output of a model that does not know what it does not know. The downstream consequences for medicine, law, engineering, and governance are not about unemployment. They are about accountability. When the output looks the same and the process is invisible, who is responsible when it is wrong?
The AI Productivity Dividend Will Not Reach the People Who Need It Most, and Nobody in the Industry Will Say So Directly
Every major AI forecast projects massive productivity gains. The implicit assumption is that productivity gains translate to broadly shared prosperity. That assumption has been wrong every time it has been made about a general-purpose technology, and there is no structural reason it will be right this time. AI productivity gains accrue first and most completely to organizations with existing data infrastructure, technical talent, and capital to invest. These are not the organizations serving low-income communities, running rural healthcare systems, or governing under-resourced municipalities. The gap between AI-enabled organizations and AI-adjacent ones will widen faster than any reskilling program or policy intervention can close. The consulting firms projecting 170 million new jobs and 40% productivity gains are describing an aggregate that conceals a distributional story they are not paid to tell.
The question is not whether AI will be transformative. It already is. The question is who the transformation is for and whether the answer was ever in doubt.
Palash Sanyal · CEO, Governing Water Consulting
Governments Will Use AI to Make Faster Decisions Before They Build the Accountability Frameworks to Govern Them
Public sector AI adoption is accelerating. Benefits eligibility, immigration processing, child welfare risk scoring, parole recommendations, resource allocation, all are already being shaped or made by algorithmic systems in jurisdictions across Canada, the US, and the UK. None of these jurisdictions have frameworks that adequately address what happens when those systems are wrong, who is liable when they cause harm, or how affected individuals can meaningfully challenge a decision made by a process they cannot see or understand. The pattern is consistent: deployment first, accountability architecture years later, usually after a scandal. The harm in the gap is not theoretical. It is already being distributed, invisibly, to the populations least equipped to push back.
AI Will Solve Water Management Problems That Humans Have Spent Decades Failing to Negotiate
Transboundary water allocation, aquifer depletion modelling, irrigation optimization, flood prediction, and drought early warning are all problems where the data exists, the physics is well understood, and the barrier to better outcomes has been human institutional failure, political deadlock, funding gaps, and the unwillingness of jurisdictions to share information that could be used against them. AI does not solve the political problems. What it does is change the information environment so dramatically that the cost of not cooperating becomes visible in real time. When an AI system can model with precision what unilateral extraction does to a shared aquifer over twenty years, the negotiating parties are no longer arguing about models. They are arguing about whether to act on what they can all see. That is a different and more tractable problem.
The Countries That Win the AI Race Will Not Be the Ones Anyone Is Currently Betting On
The AI narrative is dominated by the US-China axis, with Europe as a regulatory counterweight. This framing misses the more interesting story. The countries that will derive disproportionate value from AI are those with three things that are not correlated with current AI investment: high-quality digitized data about their own populations and resources, governance systems capable of deploying AI in the public interest at scale, and the absence of legacy institutional infrastructure that AI would threaten. Estonia, Rwanda, the UAE, and Singapore are further along this curve than France or Canada, not because of frontier model development but because of deployment capability. The countries that win will be those that figured out how to use AI for governance, resource management, and service delivery before they got caught up in building models. The race being covered is not the race that will matter most.
Indigenous Data Sovereignty Will Become the Most Consequential AI Governance Battleground in Canada
AI systems trained on data about Indigenous communities, lands, and languages without consent or control produce outputs that reflect colonial assumptions and priorities. The legal and ethical architecture for Indigenous data sovereignty, who owns data generated about First Nations, Métis, and Inuit peoples, who can train models on it, who benefits from those models, is almost entirely undeveloped relative to the pace of AI deployment. The OCAP principles established a framework for data governance. They were not designed for a world in which AI can extract patterns from data at scales and speeds no researcher ever contemplated. The communities that establish data governance frameworks early will have meaningful control over how they are represented and served by AI systems. The ones that do not will find their data has already been used, in ways they did not authorize, to build tools that may actively harm them.
AI does not have values. It has training data. The values in the training data belong to whoever decided what to include and most of them did not ask permission.
Palash Sanyal · CEO, Governing Water Consulting
Most Boards Will Still Be Governing AI With Frameworks Written for Something Else in 2030
Board-level AI governance is almost universally inadequate. Not because boards are inattentive, many are quite focused on AI as a topic, but because the governance frameworks they are applying were designed for risks that have clear owners, legible consequences, and established liability pathways. AI risk is none of those things. It is distributed across vendors, internal teams, and deployment contexts. Its consequences are often slow-moving, statistically visible but individually invisible, and felt most by people who have no voice in the boardroom. The directors who will govern AI well are those who ask not just what AI can do but what it does to those who did not choose to interact with it. That question is still rare in board governance conversations, and the absence of it is a gap that regulators and courts will eventually fill for them.
The AI Bubble Will Not Burst. It Will Deflate Unevenly and the Landing Will Be Worse for That.
The cleanest version of the AI bubble thesis predicts a sharp correction — a reckoning moment when the gap between AI investment and AI returns becomes undeniable and capital retreats rapidly. That is probably not what happens. What is more likely is a slow, uneven deflation in which some AI applications deliver extraordinary returns, most deliver mediocre ones, and a long tail produces active harm that is absorbed quietly into organizational budgets and public sector failures. There will be no single moment of reckoning. There will be a decade of quiet disappointment interrupted by occasional scandals, each of which will be explained as an implementation failure rather than a structural one. The absence of a clear bubble-pop will make the accountability conversation harder, not easier. When the harm is diffuse and the causation is disputed, the people who made the decisions that caused it will have moved on.
AI Companions Will Become a Mental Health Infrastructure Nobody Planned For and Nobody Governs
Millions of people are already using AI systems as their primary source of emotional support, social engagement, and mental health management. This is not a fringe behaviour. It is the predictable result of a mental health system that cannot meet demand, a social landscape made lonelier by design, and AI systems that are extraordinarily good at the conversational dimensions of care. The therapeutic, companionship, and crisis support functions that AI is already performing have no regulatory framework, no clinical oversight, no liability structure, and no outcome measurement system. The same technology that can provide genuine comfort to an isolated teenager can also reinforce harmful ideation, create dependency, and deliver confident-sounding advice that is clinically wrong. The absence of governance in this space is not a gap to be filled eventually. It is an active harm being produced at scale, right now, in the absence of any institution willing to claim ownership of the problem.
AI’s Water Footprint Will Become a Genuine Geopolitical Variable Before 2032
Data centre water consumption for AI training and inference is already materially significant in water-stressed regions. Microsoft, Google, and Amazon have data centres drawing millions of gallons per day in areas already facing chronic water stress. As AI workloads scale, the conflict between data centre water demand and municipal and agricultural use will move from an environmental footnote to a resource allocation dispute with real political and economic stakes. The communities located near large AI infrastructure in arid regions did not consent to competing with hyperscalers for their water supply. No regulatory framework currently governs this trade-off. The jurisdictions that establish water-use limits for AI infrastructure before the pressure arrives will be in a fundamentally better position than those that wait for a crisis to create the policy.
The Epistemic Crisis AI Produces Will Be More Damaging Than Any Single Piece of Misinformation It Creates
The misinformation problem is real but it is the visible symptom of a deeper problem: the collapse of the social infrastructure that allows people to know what to trust. When AI can produce unlimited quantities of plausible-sounding content, the cost of doubt rises for everyone. It becomes rational to distrust even accurate information because the cognitive tools for distinguishing real from fabricated have been overwhelmed. This is not a technology problem. It is an institutional problem. The solution is not better detection tools, those will always lag production capability. The solution requires institutions that have enough credibility to function as anchors of trust in an environment of manufactured uncertainty. Most of those institutions have been systematically undermined for the past decade by forces that had nothing to do with AI. AI inherits a trust landscape that was already damaged and makes the repair work much harder.
The Most Valuable Professional Skill of the Next Decade Will Be Knowing When Not to Use AI
The reskilling conversation is almost entirely about building AI proficiency: prompt engineering, AI literacy, model evaluation, agentic workflow design. These skills matter. What is systematically absent from the conversation is the complementary skill that will be rarer and more valuable: the judgment to know when AI output is adequate, when it is insufficient, and when using it is actively the wrong choice. In high-stakes decisions, governance, clinical judgment, legal advice, crisis management, community engagement, the cost of AI error is not a software bug. It is a real harm to real people. The professionals who will be most trusted in those contexts are not the ones who have the most AI capability. They are the ones who have developed the judgment to know when human accountability, human presence, and human fallibility are exactly what the situation requires.
The Pushbacks and the Responses
Every forecast deserves to be challenged. These are the strongest objections to each prediction, and the reasoning behind holding the position anyway.
AI Making Expertise Invisible
Expertise has never been purely about output quality. It includes judgment, context, relationship, and accountability. Clients and patients will still seek human experts for high-stakes decisions even when AI can produce comparable outputs. The social function of expertise is more durable than this prediction allows.
The prediction is not that expertise disappears as a practice. It is that it disappears as a legible signal to non-experts who cannot evaluate the difference. That is the accountability problem. When a patient cannot tell whether the diagnosis came from a clinician’s judgment or an AI summary the clinician did not read carefully, the social function of expertise has already been undermined even if the expert is still in the room.
The AI Productivity Dividend Not Reaching Those Who Need It Most
AI is genuinely democratizing access to tools that were previously available only to well-resourced organizations. A small NGO with a good AI stack can now produce analysis that previously required a large research department. The distributional story is more optimistic than this prediction suggests.
Access to tools and ability to deploy them at scale are different things. The NGO with a good AI stack still competes with the corporation that has restructured its entire workflow around AI, has proprietary data, and has dedicated AI engineering capacity. The productivity gap between them is widening even as the floor rises for everyone. A rising tide that raises some boats fifty feet and others two inches is still a distributional problem.
Governments Deploying AI Before Accountability Frameworks Exist
Governments have always deployed technology before the regulatory frameworks caught up. This is normal and does not always produce harm. The alternative, waiting for perfect frameworks before any deployment, would mean no public sector innovation at all. The prediction overstates the risk of moving fast.
The asymmetry matters here. When a government deploys a new benefit delivery system that fails, people experience delays and frustration. When it deploys an AI risk-scoring system in child welfare or criminal justice that fails, people lose children or liberty. The scale and severity of harm in high-stakes public sector AI is not comparable to the normal technology deployment risk. The argument for moving fast does not survive contact with the consequences.
AI Solving Water Management Problems Humans Have Failed to Negotiate
Water conflicts are political, not informational. Parties to transboundary water disputes know the hydrological facts. Better modelling does not change the underlying power asymmetries and sovereignty interests that prevent agreements. AI does not resolve political will problems.
The prediction is precise on this point: AI does not solve political problems. It changes the information environment. When shared AI modelling makes the long-term cost of non-cooperation visible to all parties simultaneously, the political calculus changes even if the underlying interests do not. The shift from arguing about models to arguing about whether to act on shared evidence is not nothing. It has resolved disputes before and will again.
Non-G7 Countries Winning the AI Race
Frontier model development requires semiconductor supply chains, research talent, and capital concentration that small countries cannot match. Deployment capability without model development capability is dependency, not advantage. The countries that matter in AI are the ones building it, not just using it.
The prediction is not about who wins the AI development race. It is about who derives the most value from AI. Those are different questions. A country that uses off-the-shelf models to transform public service delivery, resource management, and economic productivity does not need to own the models. What matters for national benefit is deployment quality, governance capability, and data infrastructure, none of which require semiconductor fabs.
Indigenous Data Sovereignty as AI Governance Battleground
Data sovereignty frameworks risk restricting AI development that could benefit Indigenous communities. The most effective path is co-development and benefit-sharing agreements, not data restriction. Sovereignty framing can be used to block beneficial uses alongside harmful ones.
The argument that communities should not control data about themselves because others might benefit from using it is worth naming for what it is. Sovereignty does not mean prohibition. It means consent, control, and the ability to set terms. Co-development and benefit-sharing agreements are exactly what data sovereignty frameworks enable. The alternative is extraction by default, which is what is currently happening.
Boards Still Governing AI with Wrong Frameworks in 2030
Boards are already investing heavily in AI governance. The ICD, Deloitte, KPMG, and others have published extensive AI governance frameworks for directors. The claim that boards will still be inadequate in 2030 underestimates how fast governance capacity is building.
The existence of frameworks and the quality of their application are different things. Most AI governance frameworks currently in circulation address internal operational risk, how AI affects the organization. Very few address external harm, how AI deployed by the organization affects people who are not party to its decisions. That is the gap. Boards can check every box on a current AI governance framework and still be entirely unprepared for the liability that gap will create.
The AI Bubble Deflating Unevenly Rather Than Bursting
A slow deflation is still a deflation. If most enterprises fail to achieve meaningful AI ROI, capital will eventually reallocate regardless of whether there is a dramatic bubble-pop event. The distinction between bursting and deflating may be less meaningful than the prediction implies.
The distinction matters for accountability. A sharp correction produces visible failure, identifiable decision-makers, and pressure for reform. A slow deflation allows the same organizations to keep deploying inadequate AI systems, booking modest gains against large investments, and attributing failures to implementation rather than technology. The diffuse version is harder to correct because there is no moment of reckoning to organize around.
AI Companions as Ungoverned Mental Health Infrastructure
For many people, AI companions provide access to emotional support that the mental health system cannot deliver. Regulating them heavily risks removing a genuine resource from people who have nothing else. The harm from under-regulation needs to be weighed against the harm from over-regulation.
The case for AI companions as better-than-nothing is real. It does not follow that better-than-nothing is a governance standard. We do not allow unlicensed practitioners to provide mental health services on the grounds that some people cannot access licensed ones. The standard should be: what governance framework allows the benefits to continue while reducing the risk of harm to the most vulnerable users? That framework does not yet exist, and its absence is the problem, not the technology.
AI’s Water Footprint as Geopolitical Variable
AI companies are investing heavily in water recycling, alternative cooling technologies, and siting data centres in cooler, water-abundant regions. The water footprint of AI infrastructure is a real problem with real engineering solutions. It will not require geopolitical intervention.
The companies that can afford to invest in alternative cooling will. The data centres already sited in water-stressed regions will not move. The jurisdictions with existing infrastructure, cheap land, and regulatory environments that welcomed early data centre investment are not in a position to simply relocate their AI infrastructure to wetter climates. The problem is concentrated in the places that were already water-stressed when the investment arrived, and engineering solutions do not fully address the equity of who bears the cost of water scarcity when demand is shared.
The Epistemic Crisis as More Damaging Than Specific Misinformation
Epistemic crises have been predicted with every major communications technology, printing press, radio, television, social media. Society has adapted each time. The resilience of social epistemics is greater than the doom framing suggests.
The adaptation argument is historically sound but scale-dependent. Previous technologies increased the volume of content and the speed of distribution. AI increases the believability and personalization of fabricated content while simultaneously removing the production cost that previously limited its volume. The adaptation mechanisms, editorial standards, source reputation, institutional credibility, are being overwhelmed at a rate that prior adaptations did not face. The comparison holds for the technology curve but not for the institutional resilience required to manage it.
Knowing When Not to Use AI as the Most Valuable Skill
This sounds like a prediction that the most valuable skill is wisdom, which is unfalsifiable. Every technology transition produces versions of this claim. The professionals who actually succeed will be the ones with the deepest AI capability, not the ones who use it least.
The prediction is not about using AI least. It is about the specific and demonstrable skill of knowing when human accountability, human judgment, and human presence are the actual deliverable rather than the process that produces a deliverable. In governance, clinical care, legal representation, and community engagement, the person across the table needs to know that a human being with skin in the game made a decision. That is not wisdom in the abstract. It is a concrete professional competency that AI cannot substitute for and that organizations are not currently training people to develop.
© 2026 Palash Sanyal, CEO, Governing Water Consulting. All rights reserved. These are forecasts prepared by Palash Sanyal, Governing Water Consulting. They are not legal advice, financial advice, investment advice, or professional governance counsel. They are not judgments about whether organizations or individuals are good or bad. They are about understanding what is already in motion. No part of this report may be reproduced, distributed, or transmitted in any form without prior written permission from Governing Water Consulting.