In this episode of Decoder, Hayden Field interviews Heidy Khlaaf, Chief AI Scientist at the AI Now Institute, about the shift in the AI industry towards military applications. They discuss the ethical implications of companies like OpenAI and Anthropic partnering with the defense sector, the potential risks of using commercial AI models in military systems, and the redefinition of AI safety standards. Khlaaf expresses concerns about the lack of rigorous testing and evaluation for AI systems used in defense, the potential for compromised data to influence AI behavior, and the dangers of focusing on hypothetical risks like CBRN weapons while neglecting current safety and security issues. She also critiques the AI industry's approach to safety, arguing that it has been co-opted and redefined to bypass traditional safety measures.
Sign in to continue reading, translating and more.
Continue