news 
Anthropic CEO says spies are after $100M AI secrets in a ‘few lines of code’
In a recent event at the Council on Foreign Relations, Anthropic’s Chief Executive Officer Dario Amodei shared his concerns regarding the potential theft of costly “algorithmic secrets” from top AI companies in the United States. He believes that these Chinese entities may be using industrial espionage to obtain valuable information.
Amodei emphasized that some of these algorithmic secrets can be worth a staggering $100 million and only comprise a few lines of code. He is convinced that some individuals are likely trying to steal this sensitive data, possibly succeeding in their efforts.
The CEO further highlighted the importance of receiving additional support from the U.S. government to protect against these risks. While he did not specify what kind of assistance would be required, it is essential for AI companies like Anthropic to collaborate with intelligence agencies and other allies.
This concern aligns with Amodei’s earlier sentiments expressed in his essay “Machines of Loving Grace,” where he emphasized the potential misuse of artificial intelligence (AI) by authoritarian countries.
