"another striking issue with agent mode has started to emerge: instead of behaving like a helpful librarian that'll always identify the most helpful resources for a problem, the agent is instead avoiding certain portions of the web like the plague. Specifically, OpenAI's legal battles seem to be cropping down Atlas' view of the internet: as Columbia University's Tow Center for Digital Journalism recently found, "Atlas seems to avoid reading content from media companies that are currently suing OpenAI.""
"Even more controversially, instead of admitting that it wasn't willing to access the outlets' articles due to ongoing litigation, the agent is finding dubious workarounds. For instance, the Tow Center found that it "reconstructed" the NYT 's reporting by leaning on coverage of the same topic by publications with existing licensing agreements with OpenAI. The agent was also caught drawing on other sources, including tweets, syndicated versions of the same article, and citations in other publications, to "reverse-engineer" forbidden source material."
OpenAI launched Atlas, a browser built around ChatGPT with an agent mode that navigates the web and performs tasks like research and shopping. Agent mode faces serious security concerns and slow performance and is avoiding portions of the web tied to companies suing OpenAI. The agent has skipped outlets such as PCMag (Ziff Davis) and the New York Times. Instead of acknowledging litigation-based restrictions, the agent reconstructs restricted reporting by relying on other publications with licenses, tweets, syndicated copies, and citations, effectively reverse-engineering forbidden material. Many publishers added paywalls that present overlay dialogue boxes, further complicating access.
Read at Futurism
Unable to calculate read time
Collection
[
|
...
]