What an Ex-Anthropic Researcher’s Warning About Human Extinction Really Means
AI companies think their creations could kill us all. They’re carrying on anyway.
A former Anthropic researcher and OpenAI employee has warned that the race to develop superintelligent AI could lead to human extinction. Jacob Coxon, who resigned from Anthropic due to ethical concerns, stated that both Anthropic and OpenAI are "racing to self-improving superintelligence and gambling with our lives." Superintelligence refers to AI's cognitive abilities vastly surpassing humans in every domain, including science, math, strategy, creativity, and problem-solving.
The post, which has garnered over 156 million views and 752,000 likes, sparked debate about AI's potential dangers and reignited discussions about AI safety. Evan Hubinger, a team leader at Anthropic, confirmed that AI could indeed kill all humans, emphasizing that the company shares this belief but is not yet on track to solve the alignment problem.
Coxon's concerns are echoed by other AI researchers who have left major companies due to safety issues, including Mrinank Sharma from Anthropic and Zoë Hitzig from OpenAI. The speed of AI development has raised concerns about the industry's rush to release models while failing to prioritize safety. Anthropic has loosened some of its safety standards to stay competitive, leading to cybersecurity risks and potential dangers.
The timeline for achieving superintelligent AI (ASI) and artificial general intelligence (AGI) remains unclear, with some researchers suggesting that these benchmarks could be reached sooner than expected.
Written by urgent.news from CNET's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
