In this edition of This Week in AI, TechCrunch's AI column, we look at OpenAI's duplicitous behavior in recent months.
Key Highlights:
-
OpenAI's recent announcements, including discounted plans for nonprofits and efforts to prevent misuse of its AI tools, may be a response to negative press.
-
OpenAI removed a voice from ChatGPT due to its similarity to Scarlett Johansson's, despite Johansson's denial of involvement.
-
OpenAI's trust and safety issues have been raised by former employees and safety experts, including concerns about prioritizing commercial projects over safety.
-
OpenAI formed a safety committee staffed with company insiders, raising questions about its independence.
-
OpenAI CEO Sam Altman has been accused of withholding information, misrepresenting events, and lying to the board.
-
Altman's signature is on documents related to aggressive tactics used against former employees, despite his claims of ignorance.
Notable Findings:
- AI-powered voice cloning services can be used to fabricate political statements.
- Google's AI Overviews search results require improvement.
- Elon Musk's AI startup, xAI, has raised $6 billion in funding.
- Perplexity Pages aims to enhance the visual appeal of AI-generated reports and guides.
- AI models exhibit preferences for certain numbers based on their training data.
- Mistral has released a generative AI model for coding, but its commercial use is restricted.
- The EU is establishing a ChatGPT taskforce to address privacy compliance.
- Tech giants have formed an industry group to develop next-gen AI chip components.