The news: As ChatGPT moves further into the mainstream and gains users, it continues to face critical examination.

• Its parent OpenAI and the Associated Press (AP) have formed a novel two-year partnership, as reported by Axios.
• Meanwhile, the Federal Trade Commission (FTC) is investigating the app for potential consumer harm through data collection and misinformation.

The deal: OpenAI will be able to access AP's news archives from 1985 to improve training of its language models. AP, in return, gains OpenAI's technological knowhow. The specific technical details are still in development.

Why it matters: This marks a significant collaboration between a US news firm and an AI company.

• OpenAI COO Brad Lightcap sees AP's input as invaluable for enhancing OpenAI's systems. AP has a history of using automation in news production, from recapping corporate earnings to local sports events.
• This partnership could help AP incorporate generative AI responsibly in future services. AP says it hasn’t used generative AI in news stories yet.

The scrutiny: News that the FTC is looking into ChatGPT for potential consumer harm through data collection and misinformation means OpenAI could be facing its first significant regulatory challenge. It’s also the US government’s most explicit attempt to regulate AI writ large.

• ChatGPT uses a neural network to learn skills by data analysis. However, this can also lead to information distortions known as hallucinations.
• OpenAI critics have lodged FTC complaints over alleged algorithmic bias, privacy issues, and hallucination by ChatGPT. OpenAI has acknowledged some product limitations, including the risk of broad discrimination.

The final word: While the FTC probe could spotlight OpenAI’s role in the AI debate, highlighting potential risks to jobs, security, and democracy, the AP-OpenAI partnership is one more sign that the cat’s out of the bag. Given consumer and business interest in generative AI, there’s no reason to assume that the technology will go away.