AI Superintelligence: Global Call for Prohibition

A global push calls for banning superintelligence until safety is guaranteed.

The Future of Life Institute launched a sweeping initiative that unites a remarkably diverse coalition of voices — world-renowned AI scientists, faith leaders, policymakers, artists and other public figures — to demand a prohibition on developing frontier AI beyond a certain level. The statement highlights Yoshua Bengio, among the world’s most cited AI researchers, and notes the signatories span Nobel Laureates, Turing Award winners, national security experts and cultural leaders. Notable signatories include national security figures such as Mike Mullen, the U.S. Navy Admiral (retired).

At the core, the coalition argues that frontier AI systems could surpass most people on many cognitive tasks within a few years, presenting a double-edged paradox: immense potential to solve global problems, paired with risks of misalignment or misuse. The group says any path toward “superintelligence” must be blocked until the technology is reliably safe, controllable, and has broad public buy-in. The call is explicit: prohibit development of superintelligence until safeguards exist and the public has a meaningful say in decisions that shape the technology’s trajectory.

The release also cites a poll indicating the public is wary of moving forward without stronger oversight. In other words, the coalition isn’t just appealing to tech elites; they’re underscoring that public legitimacy will be essential if any ban or limit is to endure beyond a few headlines. The question now becomes: how would such a prohibition be defined, policed, and sustained across borders when the frontier of AI is inherently global and fast-moving?

Policy-makers and industry observers will watch closely how a prohibition could be translated into concrete rules. The report implies a design principle: systems that are categorically incapable of harming people should be a prerequisite for any further progress, and public buy-in must precede deployment of capabilities that approach human or superhuman cognition. In practice, this raises thorny questions: which capabilities count as “superintelligence”? what thresholds trigger prohibitions? who enforces them, and how do you verify compliance across dozens of jurisdictions and research labs?

For the AI industry, the signal is as much about governance as it is about safety. A prohibition, if adopted, would likely reshape funding patterns, labor mobility, and disclosure norms. Labs chasing breakthroughs could pivot toward safety engineering, robustness testing, and transparency, while worrying startups may face higher barriers to scale. The spotlight will also fall on regulatory coordination: without a broad, credible international consensus, firms could relocate R&D to jurisdictions with laxer rules, creating a different risk profile than envisioned by the coalition.

Two practitioner takeaways stand out. First, definition drift is a real risk. Policymakers would need crisp, enforceable definitions around “superintelligence” and “near-superintelligence” to avoid loopholes that let firms edge around limits while claiming compliance. Second, enforcement would require credible, verifiable safeguards and cross-border cooperation. A prohibition that relies on voluntary compliance or ambiguous “buy-in” risks being unevenly applied and quickly undermined by cross-national competition for talent and funding.

Beyond the immediate policy question, the call places a broader debate front-and-center: should progress toward ever more capable AI be tempered by public consent and robust safety guarantees, or should it march forward with safety deep in the design? The coalition’s statement clearly leans toward first principles—safety, controllability, and public legitimacy—before unlocking deeper levels of machine intelligence. Whether policymakers worldwide translate that into durable law remains to be seen, but the rhetoric is unmistakable: the era of frontier AI doesn’t just demand better code; it demands a better consensus about what we’re building and who gets to decide.

Sources

Prominent Scientists, Faith Leaders, Policymakers and Artists Call for a Prohibition on Superintelligence, as Poll Shows Americans Don’t Want It