It’s been quite an eventful week for AI, especially OpenAI, maker of ChatGPT. The company was sued in yet another class action lawsuit for possible copyright infringement and privacy violations, and was subject to a wide-reaching Civil Investigative Demand from the FTC. To top it all off, some media articles questioned whether GPT-4 got “dumber” and “lazier.”
Biggest AI news of the week
OpenAI subject to ftc investigation
- Biggest news: The Federal Trade Commission (FTC) issued a Civil Investigative Demand (CID) to OpenAI. The CID spans 20 pages, listing 49 questions OpenAI must answer, plus 17 different items OpenAI is asked to turn over relevant documents. The FTC’s far-reaching inquiry covers two areas: OpenAI’s statements and practices regarding (1) privacy and the handling of personal data/information on ChatGPT, and (2) the potential spread of defamation and false information about people in responses by ChatGPT. [Link]
3 new proposed class action lawsuits v. OpenAI, meta, google
- Three new proposed class action lawsuits for alleged copyright infringement and/or privacy were filed against OpenAI, Meta, and Alphabet, respectively. Two suits involved Sarah Silverman and other book authors who allege their books were used to train OpenAI’s or Meta’s AI without permission based on illegal digital copies of the books from shadow libraries. [A complete list of the lawsuits]
elon musk launches new company xAI
- Elon Musk launched his own AI company called xAI. Must co-founded OpenAI, but left. [Link]
critics say gpt-4 got dumber and lazier
- If the news weren’t bad enough for OpenAI, critics claim GPT-4 is “lazier and dumber.” [Link]
openai announces license deals with ap & shutterstock
- On a positive front, OpenAI announced licensing deals with Shutterstock (6-year deal) and AP News. OpenAI gets to use Shutterstock’s image, video, and music libraries, and AP’s news database. AP is exploring ways to integrate OpenAI’s GPT in its business.