What if the loudest cheerleaders for AI were also the people who had been ignoring the biggest headaches it caused? That would be the paradox arising from new research mapping the divergent expectations of corporate leaders, investors, and the public on AI’s societal role.
The numbers tell the story. A full 93 percent of corporate executives and 80 percent of investors predict AI will deliver a net positive impact in the next five years. Only 58 percent of the public shares that enthusiasm. For business leaders, expected gains are concentrated in areas like productivity, innovation, profitability, and even shareholder return. Investors echo many of those themes but add in a sharper focus on areas such as sustainability and workforce readiness. The public is wary, however-almost half expect AI to take away jobs, a worry only one in five executives acknowledges.
That perception gap extends to environmental priorities, too. According to data from Just Capital, just 17% of corporate leaders today incorporate sustainability into AI deployment strategies, while 42% exclude it entirely. That contradicts investor sentiment-investors are the most likely group to believe AI might harm the environment, more in line with public concerns. Those concerns aren’t baseless. Computational intensity driven by AI is creating a furious build-out of hyperscale data centers, with demand for data center electricity expected to reach 7% of global consumption by 2030. In Virginia-the center of U.S. data hosting-those facilities already consume an astonishing 25% of the state’s power.
AI can optimize energy efficiency and real-time monitoring of emissions, but in return, the growth of the sector is outpacing the clean energy transition. Most new facilities today still rely on natural gas, undermining all pledges to corporate net-zero goals. Only the most progressive operators are experimenting with on-site fuel cells, AI-fueled climate control systems, and flexible power purchase agreements that reduce carbon footprint. These practices remain an exception rather than a rule.
Other areas where such consensus on importance is not translated into corporate action include training of the workforce. While 97% of investors and 90% of the public consider AI training critical, executives invest a smaller share of AI-related gains in upskilling than in either R&D or shareholder returns. This underinvestment risks leaving employees behind as AI reshapes workflows. Clearly, AI-driven process optimization requires new skill sets that cannot be improvised on the fly: data literacy, model oversight, and human-machine collaboration-across industries as diverse as manufacturing and finance. Only companies that embed structured reskilling programs into their AI strategies are better positioned to maintain productivity without exacerbating social backlash.
Governance and risk management form yet another layer of complexity. While all groups rank the safety of AI as a leading priority, they define the concept differently. The public sees equal threat in disinformation, malicious use, environmental harm, and loss of control. Executives focus more narrowly on disinformation and security breaks. Best practice frameworks now emphasize targeted governance for novel risks-like autonomous cyber-attacks or AI-enabled bioengineering-rather than overregulating so-called “boring AI,” such as productivity tools. Not only does this allow genuine capability leaps, but it also avoids stifling innovation in low-risk domains.
ESG performance viewed from the perspective of authentic AI adoption-that is, measured not by marketing claims but by operational integration-shows evidence that it enhances outcomes in both environmental and governance aspects. AI might accelerate green innovation by simulating sustainable materials, optimizing supply chains for lower emissions, and making possible digital twins to model environmental impacts before production. Additionally, it can strengthen internal controls by using machine learning to detect anomalies across enterprise systems, thus improving compliance and reducing operational risk. These capabilities are particularly potent in technology-intensive firms with absorptive capacity-the ability to integrate AI deeply into their processes.
The strategic implication for business leaders is straightforward: the very investors who are optimistic about AI’s upside are the same ones pushing for trustworthy sustainability strategies, formidable workforce development, and laser-focused risk management. To bridge the gap in public sentiment, mere trumpeting of productivity dividends would not suffice; serious action on the environmental and social dimensions of AI deployment would be needed. Those that align their AI strategies to these broader expectations are distinguished, not only by technological capability but by how well they manage the interlocking imperatives of innovation, responsibility, and trust.
