This Week in AI: Can We (and Could We Ever) Trust OpenAI?

Keeping up with the rapidly evolving field of AI can be challenging, but fear not! Here’s a roundup of recent developments and noteworthy research in the world of machine learning:

First off, OpenAI has introduced discounted plans for nonprofits and educational institutions, along with unveiling its latest efforts to prevent misuse of its AI tools. While these initiatives generally received positive feedback, some speculate they were timed to counter recent negative press about the company.

Speaking of which, OpenAI faced scrutiny over its AI-powered chatbot, ChatGPT, which used a voice that bore a striking resemblance to Scarlett Johansson’s. Johansson expressed concerns, stating she hadn’t authorized the use of her voice. Although OpenAI claimed any similarities were unintentional, questions linger about why CEO Sam Altman contacted Johansson just days before a high-profile demo featuring the voice.

Additionally, OpenAI has faced trust and safety challenges. Its Superalignment team, tasked with governing superintelligent AI, was promised significant compute resources but received only a fraction. This, along with other issues, led to the resignation of key team members.

In other news, TechCrunch is gearing up to launch an AI newsletter on June 5, providing regular updates on the latest developments in the field. Meanwhile, our semiregular AI column is now weekly, so keep an eye out for more editions covering the latest in AI research and industry news.