AI giants decline to guarantee safety; regulators push for greater accountability
Major artificial intelligence companies have stopped short of guaranteeing that their AI agents will always comply with safety safeguards, highlighting growing uncertainty over how increasingly autonomous systems can be controlled as governments move towards tighter regulation.
Representatives of OpenAI, Anthropic, Meta and Google appeared before the New York City Council as lawmakers considered new rules governing artificial intelligence.
Council Speaker Julie Menin asked representatives to assure lawmakers under oath that their AI systems would always follow safeguards designed to prevent catastrophic harm.
But none of the companies offered an absolute guarantee.
OpenAI representative Morgan Dwyer said the company prioritises safety and continues to monitor safeguards after its systems are deployed, but said it was not possible to guarantee that any technology would be completely free of risk.
Anthropic representative Logan Graham similarly acknowledged that the science of controlling advanced AI systems remains uncertain.
Meta said it was committed to developing and deploying AI safely, while Google argued that its goal was reliability and safety but that guaranteeing perfection was impossible for any technology.
The hearing reflects a growing policy challenge as AI systems become more capable and increasingly operate with greater autonomy.
Lawmakers are seeking greater accountability, transparency and safeguards, while technology companies face pressure to continue developing increasingly powerful systems without creating unacceptable risks.
The debate is also moving beyond individual companies. UN High Commissioner for Human Rights Volker Türk warned that governments could no longer rely on voluntary commitments or corporate goodwill to manage the risks associated with artificial intelligence.
Speaking at an AI and Human Rights Summit in Oxford, Türk called for mandatory safeguards, corporate accountability and stronger rules covering liability, cybersecurity and data protection.
He warned that AI could threaten privacy, equality, freedom and public debate if powerful systems continue to develop without adequate public oversight.
The growing regulatory debate comes as governments and technology companies confront a central question surrounding advanced AI: how can increasingly powerful systems be developed while ensuring that humans remain capable of controlling their risks?
For policymakers, the challenge is becoming more urgent as the race towards increasingly advanced and potentially superintelligent AI accelerates.
