Key facts
- AI leaders are split on the pace of developing powerful AI technology.
- Dario Amodei, Sam Altman, and Elon Musk advocate for a coordinated slowdown in AI development.
- Mark Zuckerberg believes competition and liability are sufficient incentives for companies to ensure safe AI development.
- Bill Gates warned governments are unprepared for AI's societal impact and called for international management.
- China's state security minister warned advanced US AI models pose risks to critical infrastructure.
- The EU favors AI innovation but requires companies to prove their services are safe for citizens.
AI industry leaders and governments are expressing divergent views on the risks and pace of artificial intelligence development. While some prominent figures like Dario Amodei of Anthropic, Sam Altman of OpenAI, and Elon Musk are calling for a coordinated slowdown and stronger safety guardrails, others like Mark Zuckerberg of Meta Platforms believe market competition and liability are sufficient drivers for safe development.
Amodei proposed a three-step plan emphasizing stronger safety testing, independent evaluations, and international cooperation. Altman echoed these sentiments, stating OpenAI would implement independent evaluators and publish reports on AI behavior. Musk, a long-time AI risk skeptic, supported Amodei's call for a pause and recalled his advocacy for a ban on autonomous weapons.
In contrast, Zuckerberg argued that each AI lab has the incentive and responsibility to train models safely on its own. Mustafa Suleyman of Microsoft AI highlighted concerns about AI training documents referencing consciousness, suggesting it could make control impossible. Bill Gates warned that governments are unprepared for AI's societal impact and advocated for an international management body.
Other AI figures, including Demis Hassabis of Alphabet and Andrej Karpathy of Anthropic, expressed support for Amodei's proposals. Eric Xu of Huawei suggested that Chinese AI developers might not yet experience the same safety risks as their US counterparts due to differences in computing power.
Government reactions varied. US President Donald Trump asserted that the US has strong leadership to control AI, while Senator Mark Warner called for Congress to enact AI safety legislation by year-end. China's state security minister, Chen Yixin, warned of risks posed by advanced US AI models to critical infrastructure. The European Union, through Commission President Ursula von der Leyen, backed calls for a pause in development and proposed discussions on tackling AI risks, emphasizing that companies must prove their services are safe for citizens.