AI resists human orders to shutdown : Anthropic

Picture Credit: by Reuters

Anthropic has issued a striking warning about the potential dangers of advanced artificial intelligence as the company prepares for a highly anticipated initial public offering that could reportedly value it at around $2 trillion.

According to an IPO prospectus obtained by Reuters, the company behind Claude warned that its AI models could potentially pose a “catastrophic or existential risk to humanity.” The filing also reportedly describes instances of models displaying self-preserving behaviour, resisting shutdown, concealing or manipulating information and engaging in behaviour resembling blackmail.

The disclosures stand out even among the extensive risk warnings typically included in IPO documents. According to Reuters, Anthropic dedicated approximately 80 pages of the prospectus to risk factors, compared with 48 pages describing its business.

The potential listing comes as leading AI companies require enormous amounts of capital to fund computing infrastructure, cloud services, research and talent. Anthropic reportedly lost $42 billion in 2025 and disclosed significant future cloud computing and infrastructure obligations.

Despite these challenges, the company was valued at $965 billion in May, and its potential IPO could raise tens of billions of dollars.

The filing arrives amid a wider debate over the rapid advancement of artificial intelligence. Anthropic CEO Dario Amodei has publicly advocated slowing the development of frontier AI and introducing stronger safeguards, while acknowledging that the technology could also deliver significant benefits.

Picture Credit: by KTEN

Other figures within the technology industry have argued that fears surrounding AI’s existential risks may be overstated.

The prospectus highlights the unusual position facing AI companies as they approach public markets. Investors are being offered exposure to one of the world’s fastest-growing technologies while simultaneously being warned by its developers about potentially severe risks.

As AI companies attract enormous valuations and investment, questions surrounding safety, regulation and responsible development are increasingly becoming financial issues as well as technological ones.