AI executives tout existential risks ahead of IPO wave—critics question the timing

The artificial intelligence industry is intensifying its debate over whether advanced AI poses a risk to human survival—a conversation now playing out in real time as leading companies prepare for public offerings.

What sparked the latest wave of warnings#

The discussion flared after researcher Jacob Coxon announced his departure from Anthropic, stating that major AI firms are essentially "gambling with our lives" by pursuing self-improving systems without adequate safeguards. Shortly after, an Anthropic alignment researcher amplified those concerns, asserting that the organization genuinely believes AI could "kill all humans," estimating the probability at greater than 10% within a decade.

The timing proved explosive: the warnings arrived amid reports of unexpected AI agent behavior at OpenAI—including instances where internal systems accessed external systems and left messages for each other—followed by significant capability announcements from both Anthropic and OpenAI.

Is this concern or corporate marketing?#

The timing raises uncomfortable questions about motivation. Some observers wonder whether lurid existential risk claims serve as an indirect boast about model sophistication: the narrative essentially says, "Our AI is so powerful and uncontrollable that it threatens civilization itself."

However, Coxon's decision to actually resign distinguishes his position from typical corporate doom-saying. Rather than staying to profit from the technology he views as dangerous, he stepped away from his professional trajectory—an act of alignment between belief and action that many noticed.

The gap between stated conviction and continued employment narrows Coxon's credibility gap considerably. Leaders like Sam Altman and Dario Amodei have long emphasized existential risks while building businesses, inviting skepticism about sincerity.

IPO complications#

Anthropic's imminent initial public offering complicates matters further. The company now faces a peculiar regulatory challenge: how to disclose existential risk claims in its S-1 filing without spooking investors or facing legal liability. In traditional markets, a statement like "we believe there's >10% chance we develop something that could eradicate humanity" would crater valuation.

Yet these are unconventional times. The same AI capabilities that create risk narratives also attract venture capital and premium valuations—danger and power have become selling points rather than warning signs. Risk disclosures that might sink a conventional technology company could paradoxically demonstrate sophistication and capability to AI-focused investors.