Florida’s novel lawsuit against Roku under its privacy law has been noticed by lawyers and industry insiders.
A new preprint suggests that AIs are trained to hallucinate through their training that rewards confidence and conversely disincentivizes “I don’t know” responses.
A new paper discusses a layer in AI industry that’s frequently not talked about: the human labor that goes into “collect[ing] and annotat[ing] data, monitor[ing] and maintain[ing] algorithmic systems, keep[ing] data centers running, and min[ing] rare earth minerals—not to mention the artists, translators, writers, and actors whose work fuels so-called generative AI”
A recent audit found that the continuing budget and staffing cuts at the CFPB has left major data security risks.
A network of global privacy regulators announced an enforcement sweep into digital services’ use of underage users’ data.
The Fifth Circuit heard a case, Computer & Comm. Ind. Ass’n v. Paxton, regarding Texas’ law that would require content filtering for minors, although it seemed wary of deciding it directly instead of remanding to the District Court.
A new bill introduced in the senate, the GUARD Act, would regulate the use of chatbots by minors.
The FCC will vote later this month to reverse a Biden-era policy that added cybersecurity requirements.
OpenAI has updated its terms of service to say its models cannot be used to provide legal or medical advice. OpenAI disclaimed this as “not a new change to our terms.”
More than a dozen states have filed a motion to submit an amicus brief in Huiskamp v. ZoomInfo Tech. LLC, arguing that selling peoples’ phone numbers should be treated as commercial speech.
(compiled by Tobit Glenhaber)