Is This What We Want? is a public statement released in 2025, signed by more than 1,000 researchers, engineers, and technologists in the field of Artificial intelligence. The declaration calls for immediate and enforceable international regulation of AI development, citing concerns about unmitigated risks to societal safety, democratic institutions, and global equity. Its release marked a rare instance of collective advocacy from within the technical community, contrasting with prior industry-led calls for voluntary self-governance.
The statement emerged from a working group that included signatories from major research institutions and private companies, though the organizers chose to emphasize its grassroots character over any single organizational affiliation. The document’s title - a rhetorical question posed to policymakers and the public - reflects a deliberate appeal to shared values rather than technical expertise alone. Within weeks of publication, it had been covered by major international media and debated in parliamentary and congressional settings, but it did not lead to any immediate formal treaty or legislative change.
Background and context
The declaration arose against a backdrop of rapid commercialization of Large language models and Generative AI systems. In the years preceding its release, capabilities in Machine learning and Deep learning had advanced sharply, leading to both public enthusiasm and widespread anxiety about job displacement, misinformation, and surveillance. Earlier notable interventions - such as open letters from various tech leaders - had often been criticized as vague or self-interested. The organizers of Is This What We Want? sought to distinguish their effort by grounding it in specific evidence from recent failures and near-misses, including documented cases of autonomous systems causing harm.
Key demands and principles
The statement advances several actionable demands, rather than a generalized call for caution. Among its principal points are:
- Mandatory safety standards - requiring stress-testing of high-capability Neural networks before deployment, akin to aerospace or pharmaceutical certification.
- Transparency obligations - compelling developers to disclose training data sources, energy consumption, and known failure modes.
- Redress mechanisms - establishing legal liability for harms caused by AI systems, with a clear chain of responsibility.
- International coordination - proposing a treaty body with enforcement powers, modeled on existing nuclear or chemical weapons regimes.
The text explicitly rejects the argument that regulation would slow innovation, citing historical examples where safety rules fostered rather than hindered technological progress. It also questions the adequacy of current internal governance at firms like OpenAI, Anthropic, and Google DeepMind, arguing that profit incentives inevitably distort self-assessment.
Signatories and support
The signatory list includes prominent figures such as Joshua Tenenbaum (MIT), Melanie Mitchell (Portland State University), BAIR (Berkeley AI Research) faculty, and former industry researchers from Xerox PARC and Nokia Bell Labs. Approximately 70% of signatories were affiliated with academic institutions, while the remainder came from private companies, non-profits, and government laboratories. Not a single CEO of a major AI firm signed, though a number of mid-level engineers and product managers did, some anonymously to avoid professional reprisal. The organizers later published a methodology note confirming that signatures were verified against institutional email addresses but were otherwise self-reported.
Media and political reception
Reaction was polarized. Supportive commentators, including several members of the European Parliament and civil society groups, praised the declaration for its concrete proposals and refusal to defer to corporate narratives. Skeptics, notably industry executives and some libertarian economists, dismissed it as alarmist and argued that existing regulations were sufficient. In the United States, a congressional subcommittee hearing on AI governance referenced the document directly, though no bill was introduced. In contrast, the Bhabha Atomic Research Centre and certain national security bodies in other countries privately expressed interest in its provisions, viewing AI as a dual-use technology needing oversight.
The declaration also sparked internal debates within the Machine learning community. Some academics worried that its tone might alienate industry partners, while others countered that neutrality was no longer tenable given accelerating deployment. A counter-petition, titled 'Our Future Is Better With AI,' gathered fewer than 200 signatures within the same period, suggesting significant intellectual weight behind the original statement.
Impact and legacy
As of late 2025, no direct policy changes have been enacted as a result of the declaration. However, it has become a common reference point in subsequent proposals, including revisions to the Artificial intelligence impact assessments used by Waymo and other safety-critical applications. Several university courses now use the text as a case study in technology governance. In 2026, an annual review conference was established to track progress on its demands, though participation from the largest companies remains sparse. The long-term significance of Is This What We Want? may lie less in immediate legislative effect than in normalizing the idea that technical experts have both the right and responsibility to speak openly on matters of public welfare.