As Claude disrupts the stock market, Anthropic researchers warn "The world is in danger"

Investing.com - If you’ve been paying close attention to the stock market lately, a pattern has become impossible to ignore: almost every sector is now feeling the threat brought by artificial intelligence. This disruption initially started in the software industry but has quickly spread outward—impacting brokerages, insurance companies, real estate, customer service, education, and more.

In the past few weeks, nearly every trading day has seen clear sell-offs related to fears of AI replacement. Investors are no longer asking whether industries will be affected by AI, but rather how quickly and deeply the impact will occur.

At the center of this anxiety is Claude, a large language model developed by AI startup Anthropic. Claude has been praised for its reasoning ability and rapid improvement cycle, and it is increasingly seen not just as a productivity tool but as a direct threat to entire business models.

However, as Claude’s capabilities continue to shake up the market, a more alarming signal has emerged from within Anthropic.

This week, the company’s PhD researcher Mrinank Sharma announced his resignation on X— and he did not leave quietly.

“Today is my last day at Anthropic,” Sharma wrote, along with a disturbing assessment: “The world is in danger.”

Notably, Sharma cautiously stated that AI itself is not the only danger. Instead, he described what he calls a “multiple crisis”—a series of overlapping threats unfolding simultaneously.

“AI, biological weapons, and other risks are not isolated,” Sharma explained. “They are part of a series of interconnected crises currently unfolding.”

In his view, humanity is approaching a dangerous tipping point—where the rapid development of technological power outpaces society’s ability to responsibly manage these technologies.

“We seem to be approaching a critical threshold where our wisdom must grow as fast as our capacity to influence the world,” Sharma wrote. “Otherwise, we will face consequences.”

Perhaps most striking is his reflection on internal challenges working at the forefront of AI development. Despite Anthropic’s public emphasis on safety and alignment, Sharma indicated that he has repeatedly witnessed the difficulty of ensuring that values truly guide decision-making at both individual and organizational levels.

“Throughout my time here, I have repeatedly seen how difficult it is to make our values genuinely govern our actions,” he wrote. “I see this in myself, and in the organization. We are constantly under pressure to set aside the most important things—and this is true in society at large.”

Sharma’s departure comes at a time when confidence in AI governance is already fragile. Governments worldwide are rushing to regulate systems they barely understand, companies are deploying increasingly powerful models, and financial markets are struggling to price this rapid upheaval.

For investors, Claude represents efficiency, automation, and profit margin expansion—but Sharma’s warning underscores a deeper concern: the same forces driving explosive innovation may also be eroding the safeguards meant to keep it under control.

As AI continues to reshape industries and disrupt markets, Sharma’s resignation raises an unsettling question: if internal researchers are so cautious, are the rest of us moving too fast to notice what’s being left behind?

This article was translated with the assistance of AI. For more information, please see our Terms of Use.