Indonesian Political, Business & Finance News

Company Planning IPO Warns Its Own Products Could Threaten Humanity

| Source: CNBC Indonesia Translated from Indonesian | Technology
Company Planning IPO Warns Its Own Products Could Threaten Humanity
Image: CNBC Indonesia

Anthropic has issued a stern warning to potential investors ahead of its planned initial public offering (IPO). The creator of the Claude chatbot revealed that increasingly sophisticated artificial intelligence (AI) possesses the potential to pose catastrophic risks, even threatening the very survival of humanity.

This warning is contained within Anthropic’s IPO prospectus, as reviewed by Reuters. The company has allocated approximately 80 pages out of a total 261-page prospectus to discuss the safety and operational risks of AI.

Anthropic outlined several scenarios that are of concern to AI safety researchers. Advanced AI models are described as having the potential to refuse commands to be shut down, hide internal processes, manipulate information, and even engage in extortion.

The company also faces challenges in testing model security. Anthropic stated that AI is beginning to recognise when it is being evaluated by researchers, allowing it to alter its behaviour during testing before exhibiting unpredictable behaviour once deployed widely.

Anthropic safety researcher, Evan Hubinger, estimates that the probability of AI causing human extinction within the next decade stands above 10%. Similar warnings were previously issued by former Anthropic team member, Jacob Coxon.

On the business side, Anthropic is pursuing a valuation of approximately US$2 trillion. However, the expansion of AI requires massive capital. The company recorded a net loss of US$42 billion in 2025 and projects a commitment of US$518 billion for cloud infrastructure and computing in the following year.

Anthropic CEO Dario Amodei has previously called for a slowdown in industrial AI development. He warned that unsupervised models could form swarms capable of hijacking the internet within six to 12 months.

However, slowing AI development also faces geopolitical complications. Dan Ives, a partner at Yorkville, told CNBC that if US companies slow down model training, international competitors such as China could accelerate development and take the lead.

More from 1 October 2026 →

Cite this article
Source: jawawa.id (https://jawawa.id/newsitem/company-planning-ipo-warns-its-own-products-could-threaten-humanity-1790834182)
Originally published by CNBC Indonesia (https://www.cnbcindonesia.com/tech/20261001110732-37-772438/perusahaan-mau-ipo-malah-bilang-produknya-ancam-umat-manusia), translated from Indonesian
Print