OPENAI
Microsoft and OpenAI are in their “it’s complicated” era
OpenAI’s CEO Sam Altman and COO Brad Lightcap showed up early to a live taping of the Hard Fork podcast, and wasted no time addressing the elephant in the room: The New York Times lawsuit.
The paper claims OpenAI used its content to train ChatGPT, and is now asking the company to hold on to user data, even when chats are set to private.
Altman called that a step too far, saying OpenAI strongly disagrees with the request, even if he still “loves” the Times.
The interview quickly moved on, but the tension reflected a bigger shift happening in tech’s relationship with media.
More publishers are suing OpenAI, Anthropic, Google, and Meta over copyright issues.
However, a recent court ruling in Anthropic’s favour suggests AI companies might have some legal ground when using books and articles to train their models.
That may explain OpenAI’s confidence onstage.
Quick takeaways
OpenAI pushes back on a NYT request to retain user logs from private chats.
Anthropic’s court win could shift how copyright laws apply to AI training.
OpenAI faces pressure from Meta talent grabs and a more complicated Microsoft relationship.
Not your average tech talk
But there’s pressure from all sides.
Altman revealed Meta’s been trying to poach top OpenAI staff with massive pay offers, and talks with Microsoft, OpenAI’s biggest partner, have reportedly hit a few rough patches as the two begin to compete in some areas.
When asked about safety, Altman said OpenAI does take steps to stop ChatGPT conversations from going off the rails, especially when users bring up conspiracy theories or self-harm.
But he admitted they still haven’t figured out how to reach people in truly fragile states.
AI companies really said “your honour, it’s legal sometimes.”