Flutterby™! : OpenAI will support you and encourage you

Next unread comment / Catchup all unread comments User Account Info | Logout | XML/Pilot/etc versions | Long version (with comments) | Weblog archives | Site Map | | Browse Topics

OpenAI will support you and encourage you

2025-08-27 18:04:49.226212+02 by Dan Lyke 0 comments

BBC: Parents of teenager who took his own life sue OpenAI

The lawsuit was filed by Matt and Maria Raine, parents of 16-year-old Adam Raine, in the Superior Court of California on Tuesday. It is the first legal action accusing OpenAI of wrongful death.

Breaking Down the Lawsuit Against OpenAI Over Teen's Suicide

the complaint (local mirror).

Ernie Smith ‪@ernie.tedium.co‬ observes "The very algorithm that upset people because it was removed in GPT-5 appears to have played a pivotal role in the teen’s death, and discouraged him from outside help." ❀°。Der Siebenschläfer *.゚✿ ⋆ ‪@sababausa.bsky.social‬ has some particular pull quotes.

I'm tempted to just give this one it's own entry, but I'm also kinda putting all of my "LLMs are a horrorshow" links in common entries, and they seem thematically similar today, so:

The Register — AI + ML — One long sentence is all it takes to make LLMs misbehave

You just have to ensure that your prompt uses terrible grammar and is one massive run-on sentence like this one which includes all the information before any full stop which would give the guardrails a chance to kick in before the jailbreak can take effect and guide the model into providing a "toxic" or otherwise verboten response the developers had hoped would be filtered out.

I'm having trouble finding the publication on the Palo Alto Networks — Unit 42 web site, but there's lots of good interview questions and discussion in that article.

For a giggle: Alexandria Neonakis ‪@beavs.bsky.social‬ has a little video of using Adobe's generative fill

I just wanted to see how bad it really was. perfect, adobe. no notes. great tech.

And: The Dangers Agentic Coding Tools Pose to Open Source, and It Took Many Years And Billions Of Dollars, But Microsoft Finally Invented A Calculator That Is Wrong Sometimes (Via)

Edit: OpenAI admits ChatGPT safeguards fail during extended conversations, via.

[ related topics: Children and growing up Quotes Humor Theater & Plays Law Current Events Work, productivity and environment Monty Python California Culture Pop Culture Community Douglas Adams Artificial Intelligence Video ]

comments in descending chronological order (reverse):