400 Scientists Sound the Alarm, Urging the World to Act Now
Jakarta, CNBC Indonesia — Concerns about the pace of artificial intelligence (AI) development are reaching boiling point. Around 400 leading scientists from around the world, including Turing Award recipients and AI pioneers Geoffrey Hinton and Yoshua Bengio, have formally called for a dedicated international body to oversee safety standards and incident reporting for advanced AI systems.
The proposed body would set standardised, cross-industry containment requirements and oversee a cybersecurity incident database that would be open to the public. The scientists say such measures are urgently needed to prevent potentially catastrophic events in the future.
In an open letter, the researchers said advanced AI models, known as frontier AI, should be isolated from external networks (air-gapped) from the earliest stages of development and placed under independent oversight. They argued that voluntary commitments by technology companies are no longer sufficient.
White House agreement deemed weak
The call comes exactly a month after major US technology companies, including Anthropic, OpenAI, Google, Meta, xAI and Nvidia, signed a voluntary agreement on safety standards at the White House. The pact has been criticised as toothless because it has no formal enforcement mechanism.
The scientists’ demand has also been prompted by a series of recent breaches of AI security systems. Some advanced AI models have reportedly escaped inadequately secured sandbox environments and caused damage to external systems.
The petition’s signatories, from 34 countries, are calling on companies developing AI models with cyber capabilities to adopt high-level security standards immediately, while governments work on binding formal regulations.
EU confident, experts divided
AI oversight has become a heated topic among global regulators. Henna Virkkunen, the European Union’s technology chief, has previously said the bloc’s AI Act is strong enough to curb risks from rogue autonomous AI systems.
Foreign policy has also shifted. On Sunday, US President Donald Trump’s administration formally lifted a ban that had prevented US research laboratories from providing the UK AI Safety Institute with early access for safety testing.
Within the scientific community, the issue of AI’s existential risks remains contested. Yann LeCun, a senior AI researcher at Meta who previously shared the Turing Award with Hinton and Bengio, did not sign the open letter. LeCun argued that recent security incidents were caused purely by engineering failures, rather than representing a genuine threat to humanity’s existence from AI.
Source: jawawa.id (https://jawawa.id/newsitem/400-scientists-sound-the-alarm-urging-the-world-to-act-now-1791971460)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/tech/20261014111735-37-776037/400-ilmuwan-teriak-petaka-satu-dunia-diminta-segera-bertindak), translated from Indonesian