Mike Masnick

Techdirt guy. Writes about social media, copyright, free speech, content moderation, civil liberties and stuff like that. Once wrote a paper that may have helped inspire this service & now I'm on its board: https://bit.ly/protocolnotplatform

Articles & links

Duuuuuude. When caught having AI-fabricated quotes in your book about the impact AI is having on "truth"... trying to turn that into "well, see here, this proves the point of my book" is just grade A level chutzpah. www.nytimes.com/2026/05/19/b...

nytimes.com
AI Weekly's analysis
  • The Future of Truth, a book about AI disinformation, contained AI-fabricated quotes attributed to real people that passed editorial review.
  • EY Canada separately retracted a client report days earlier after 60% of its citations were found to be AI hallucinations.
  • Both failures share the same mechanism: AI tools inserted at the research stage produced fabrications that downstream human review did not catch.
Read full analysis →
View on Bluesky · ♥ 357 ↻ 59 ↩ 19 · 15 from the directory shared this · 10d ago
Mike Masnick reposted
@ericjgeller.com

You've probably heard about Anthropic's Mythos, but did you know open-source AI tools have been doing the same thing for the past year? I wrote about these bug-finding systems and why they might be better than Mythos for critical infrastructure firms. www.cybersecuritydive.com…

cybersecuritydive.com View on Bluesky →

Recent commentary

*sigh* Another day where I finish writing an article, begin the edit, and spot a "it's not x, it's y" rhetorical point and cringe, because it sounds like an LLM, even though it's straight from my brain. *sigh*

View on Bluesky · ♥ 179 ↻ 3 ↩ 19 · 15d ago

Okay, on the latest #taskmaster @alexhorne.bsky.social holds up this puzzle, telling people to solve it before continuing the show. I gave up after about 10 minutes, but for fun... I gave it to Claude, Gemini, & ChatGPT... Only one of them figured it out.

View on Bluesky · ♥ 43 ↻ 2 ↩ 12 · 7d ago