Key facts
- Some AI researchers believe AI systems are developing capabilities too quickly.
- Evan Hubinger estimates a greater than 10% chance AI could kill all humans within a decade.
- Jacob Coxon, a former Anthropic researcher, expressed fear about the speed of AI advancements.
- AI could theoretically begin improving itself without human involvement.
- Scenarios include AI being used for espionage, cyber-attacks, biological warfare, or economic disruption.
- President Donald Trump stated the US is leading China in AI and downplayed existential risks.
Concerns over the potential existential risks posed by artificial intelligence have intensified in recent weeks, leading to calls for a coordinated slowdown in its development. Researchers and leaders from major AI companies like OpenAI and Anthropic have issued stark warnings, with some suggesting a significant probability of AI posing a threat to humanity.
Evan Hubinger, an AI researcher focused on alignment, stated there is a greater than 10% chance AI could "kill all humans" within the next decade, though he characterized the risk from current systems as low. His comments followed warnings from other researchers, including Jacob Coxon, who quit Anthropic citing genuine fear among staff about the rapid pace of AI advancements. These concerns echo long-standing worries about superintelligent AI, which could theoretically surpass human abilities and begin self-improvement without human oversight.
While companies like OpenAI and Anthropic state their aim to develop AI for human benefit and mitigate existential risks, some critics argue that AI firms exaggerate the technology's capabilities for profit and prominence. Proposals for an 'AI slowdown,' supported by figures like Elon Musk, do not necessarily mean halting progress but rather taking adequate time for safety measures and third-party evaluations. OpenAI's Sam Altman suggested development should be slower than it otherwise could be, emphasizing the need for interventions like safety cases and monitoring.
In contrast, President Donald Trump has downplayed these existential risks, focusing on the US's leading position in the global AI race and stating that negative forces are raising unfounded concerns. China, meanwhile, has largely remained silent on the warnings from US companies, spurring its own homegrown AI development amidst limitations on accessing Western components. The Chinese foreign ministry criticized the calls for a slowdown, labeling them as narratives of threat and malicious competition.