- AI firm Anthropic has warned that its increasingly powerful systems could pose “catastrophic or existential” risks to humanity in documents published ahead of its initial public offering.
- The filings highlight potential threats including the development of biological and chemical weapons, alongside unpredictable autonomous behaviours that could cause unintended harm.
- Anthropic devoted nearly a quarter of the documents to warnings, including the unknown long-term impacts of regular AI use on human mental health and cognitive development.
- Alongside safety warnings, the company claimed AI could dramatically accelerate scientific research and potentially help cure major diseases such as cancer and Alzheimer's.
- The stock market listing will raise capital to fund costly computing infrastructure and data centres, whilst allowing existing investors Amazon and Google to sell shares.
IN FULL
