Artificial intelligence is rapidly becoming the defining technology layer of modern business. Investment is accelerating, use cases are multiplying, and executive expectations remain extraordinarily high. Yet beneath the excitement surrounding generative AI lies a more uncomfortable reality: AI is not creating equal outcomes across organizations. In many cases, it is amplifying the gap between those already capable of navigating complexity and those struggling to do so.
A recent MIT Sloan Management Review (it will be missed!) article captured this dynamic through a fascinating field experiment with entrepreneurs in Kenya. Researchers gave hundreds of small business owners access to GPT-4 through WhatsApp, positioning it as an AI business adviser. The result was striking. Some entrepreneurs significantly improved performance, while others became worse off. The differentiator was not access to technology. It was judgment: the ability to distinguish useful AI guidance from poor recommendations and to apply insights contextually rather than blindly [1].
This insight has implications far beyond small businesses in emerging markets. In fact, it may become one of the defining leadership questions of the AI era.
For years, the dominant narrative around AI has focused on access. The assumption was that once organizations democratized AI tools, productivity and innovation would naturally follow. But the emerging reality is more nuanced. AI does not eliminate the need for human judgment. It raises the premium on it.
The strongest organizations are not necessarily those with the most advanced models or the largest budgets (although they help, I know). They are the ones capable of integrating AI into coherent decision-making systems. They understand where AI adds value, where it introduces risk, and where human intervention remains essential. Weak organizations, by contrast, often use AI to reinforce fragmentation, funcional silos, accelerate poor decisions, or create the illusion of sophistication without strengthening underlying capabilities as I discussed in my previous article.
This is why AI is increasingly functioning less as a technology advantage and more as an organizational mirror.
“AI does not replace weak leadership. It exposes it.”
In many companies today, AI adoption is happening faster than organizational maturity. Teams deploy copilots, automate workflows, and generate content at scale, while governance models, incentive systems, and operating structures remain largely unchanged. The result is often a strange paradox: organizations become more efficient in data producing while simultaneously becoming less coherent.
This pattern is especially visible in innovation and sustainability functions. Across industries, companies are enthusiastically deploying AI to accelerate reporting, automate ESG disclosures, optimize supply chains, or improve forecasting. These are valuable applications. But in too many cases, the technology is layered onto operating models that were already fragmented, reactive, or strategically inconsistent. So that creates a gap from strategy to execution.
A sustainability team may use AI to generate faster climate scenario analyses while the organization still incentivizes short-term procurement decisions that undermine long-term resilience. Innovation teams may deploy generative AI to accelerate idea generation while lacking governance systems capable of prioritizing which ideas actually matter or lacking the supply infrastructure to propel scale. The technology accelerates activity, but not necessarily transformation.
This mirrors a broader challenge already visible across corporate transformation agendas. Organizations often mistake automation for reinvention. They optimize existing systems, often not the ideal ones, rather than redesigning them.
McKinsey & Company’s latest AI research reinforces this tension. While AI adoption has become widespread, only a minority of companies are capturing meaningful enterprise-wide value from it [2]. Similarly, Boston Consulting Group (BCG)’s 2025 analysis on corporate reinvention found that companies generating disproportionate value from AI are not simply deploying more tools. They are redesigning workflows, governance, talent models, and decision processes simultaneously [3].
In other words, the advantage comes from institutional adaptability, rather than from AI alone.
This is where the MIT Kenya experiment becomes especially profound. The businesses that benefited most from AI were not necessarily the most resource-rich. They were the ones capable of applying discernment. They questioned outputs, contextualized recommendations, and integrated AI guidance with practical experience and local realities at the right points of their governance. Those who accepted recommendations too literally often made worse decisions.
The lesson for executives is uncomfortable but important: AI may increase the penalty for weak judgment.
“Generative AI scales intelligence. But it also scales confusion, bias, and poor decision-making when organizations lack the capability to absorb and question it wisely.”
This challenge is becoming increasingly relevant in a digital environment already overwhelmed by what LinkedIn’s editorial leadership recently described as “AI slop”: polished but generic content that lacks authentic perspective, expertise, or lived experience. My experience here is certainly filled with mixed feelings, excitement for what AI has to offers and bored by the same content being reposted with similar generic tones. In an AI-saturated context we are facing, originality, contextual thinking, and authentic expertise become more valuable, not less.
The same principle applies inside organizations.
Companies that rely on AI to mass-produce strategy decks, innovation concepts (have seen many beautiful idea templates), or sustainability narratives (and believe me, some can be super compelling) without strengthening critical thinking capabilities may inadvertently weaken the very muscles they need most. When everyone has access to the same models, differentiation no longer comes from generating more outputs. It comes from knowing which outputs matter.
This has enormous implications for business leadership.
For decades, managerial success often depended on information access and analytical capability. AI increasingly commoditizes both. The emerging differentiator is judgment under complexity: the ability to navigate ambiguity, integrate competing variables, recognize unintended consequences, and make coherent decisions across fragmented systems. That is why I believe, those working in sustainable innovation have a leg on it, as they already managed complex ambiguous systems.
That capability cannot be outsourced to a language model.
Harvard Business School research has already started exploring how AI changes managerial cognition and expertise, suggesting that while AI can improve execution quality, overreliance may reduce independent problem-solving capability over time [4]. MIT Sloan School of Management researchers similarly argue that organizations must avoid confusing assistance with understanding [5]. Faster answers do not necessarily produce better thinking.
This distinction matters because many organizations are entering what might become the next phase of AI adoption: operational dependence. AI is no longer experimental. It is increasingly embedded into pricing, procurement, hiring, forecasting, customer service, R&D, and strategic planning. As dependency increases, the cost of weak judgment compounds. And it compounds fast.
The implications for sustainability are equally significant. AI is already reshaping energy systems, resource optimization, climate modeling, and supply chain visibility. But without coherent governance, AI can just as easily intensify environmental pressures through rising energy consumption, accelerated consumerism, and optimization models focused exclusively on short-term efficiency. I really hope my fellow sustainability leaders engage critically with AI strategy, otherwise they may find themselves reacting to consequences rather than shaping outcomes. I explored some of the AI-sustainability paradox in a previous article of The Transformation Edge.
This is why the next generation of transformation leaders will need a different capability profile. Technical literacy will matter, but not as much as systemic thinking. Being exposed to different categories, regions, industries become an asset in this race. Success in this AI era is not measured by having the loudest adopters (and there are many here on LinkedIn), but those most capable of combining technological acceleration with human discernment, organizational coherence, and long-term judgment.
The MIT experiment ultimately reveals something deeply important about the future of leadership. AI is not leveling the playing field. It is amplifying the differences that already exist between organizations and leaders.
Strong organizations become more adaptive. Weak organizations become more exposed.
And as AI becomes embedded into every layer of decision-making, the real competitive advantage may no longer be intelligence itself.
It may be judgment. Don’t you think?
References
[1] Otis, Nicholas, Rowan Clarke, Solène Delecourt, David Holtz, and Rembrand Koning. “How AI Helps the Best and Hurts the Rest.” MIT Sloan Management Review, April 20, 2026. https://sloanreview.mit.edu/article/how-ai-helps-the-best-and-hurts-the-rest/
[2] McKinsey & Company. “The State of AI: Global Survey 2025.” https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
[3] Boston Consulting Group. “Corporate Reinvention in the Age of AI.” 2025.
[4] Harvard Business School. Research on Generative AI, Managerial Cognition, and Decision-Making, 2025.
[5] MIT Sloan Management Review. Research on AI, Organizational Learning, and Decision Quality, 2025.
