Anthropic CEO Dario Amodei highlighted the risks and the potential benefits of artificial intelligence, with the fact that there is a 25 percent chance that AI could lead to catastrophic outcomes, but a 75 percent chance that it could produce many positive results on the Axios AI + DC summit. Amodei explained that, through ‘really, really bad’, he did not refer to minor issues such as autocorrelic errors, but to scenarios that can destabilize the societal systems, create existential threats or lead to serious abuse of AI. He previously warned that AI could replace up to half of all access-level-white collar posts and expressed concern over the export of advanced chips to China, indicating the economic and geopolitical interests of AI advancement. By probably drawing up the discussion, AMODEI emphasizes that although risks exist, human choices and management can influence the outcomes, which provide a balanced perspective that recognizes the danger, but also potential opportunities. Meanwhile, earlier this month, it was reported that Anthropic PBC agreed to a minimum payout of $ 1.5 billion, plus interest, to resolve a class action case that accused the AI commiters of illegally scrapping millions of copyright books to train his language models. Some of the largest AI-related settlements filed in a federal court in San Francisco on Friday are the settlement as one of the biggest agreements to date with AI and intellectual property. A hearing was scheduled for December, with plaintiffs sought damages that could have reached trillions of dollars, possibly jeopardizing the future of the company. Compensation for millions of writers The settlements cover claims of up to seven million writers, with initial payouts expected at around $ 3,000 per book for an initial set of 500,000 works. If additional claims are approved, total compensation may increase further. Anthropic also promised to remove the disputed information allegedly acquired illegally. Legal representatives for the writers described the settlement as unprecedented. Justin Nelson of Susman Godfrey said: “This agreement sends a strong message to AI businesses and content creators that the use of copyright material from pirate sources is unacceptable.” Although the financial impact is essential, legal analysts have noted that anthropically can avoid a much worse result. Tod Cohen, partner at Manatt, Phelps & Phillips, said: “The settlement represents less than one percent of the company’s valuation. It appears to be a favorable result for anthropic because it avoids dissolving its models and only needs to remove the disputed data.”
Dario Amodei, CEO of Anthropic: 25% Chance AI can be catastrophic, 75% chance of it to flourish
