Technology

"When we detect users who are planning to harm others, we route their conversations to specialized pipelines where they are reviewed by a small team trained on our usage policies and who are authorized to take action, including banning accounts," the blog post notes. "If human reviewers determine that a case involves an imminent threat of serious physical harm to others, we may refer it to law enforcement."

See? Even the people who make AI don't trust it with important decisions. And the "trained" humans don't even see it if the AI doesn't flag it first. This is just a microcosm of why AI is always the weakest link in any workflow.

This is exactly the use-case for an LLM and even OpenAI can't make it work.

Of course they are. They are a literal data farm. People need to stop using it.

Funny, now you'll have the cops arresting you for prompts like "how to survive being homeless?", rather than social services when you prompt "how to avoid being homeless?".

And will authorities be called when someone prompts "how to shoot wild animals?" when asking about wildlife photography? 😆

"Hey ChatGPT, how many human corpses can 12 pigs who haven't been fed in a week process"?

Well if idiots share their crime plans with a corpoo, what do they expect.

However, police won't do anything.

But even if they do, prosecutors won't.

Just look at that incident in Nevada with Israeli pedophile.

Cops bust him, federal prosecutor refused to charge him..

Laws are enforced selectively

Lazy authors of crime themed novels are sweating so heavily right now.

Yo, I was just joking about making a gallon of PCP

I’m cool with this! OpenAI ought to send those suicide instructions it created for that teen and send it to those pigs for some inspo :)

I gotta say... imagine being the police department on the receiving end of that firehose.

This sounds like PR deflection from egging on the kid to suicide. But I still don't doubt it.

I asked chatgpt some spicy questions a long time ago (on the order of months), and it was about privacy questions and maintaining privacy online. I even asked it how to keep private from itself. I deleted that account many months ago and it was not immediately linked to my 'real' identity. I've also asked it some other spicy questions about stuff I won't reveal here. Again. Months ago, and by it's own admission if something is reported, it will be acted upon almost immediately. If something is super illegal (as in, I want to kill so-and-so individual) it not only reports it for immediate review but also reports the relevant law enforcement right away.

Chatgpt told me this.

The only way to use chatgpt if you must is with A: a throwaway email (must be reusable) such as those on darknet emails or something that doesnt link it to you by name (such as proton or tutamail). B: over the Tor network.

Self-harm doesn't count apparently

Can call the police when you threaten to hurt others, actively prevents teens from seeking outside help when they talk about killing themselves.

All of this is so fucking bizarre I can't even wrap my head around it anymore. It's a bot. How the fuck is it suddenly killing people? How is talking to a bot a crime now? Did everyone lose their minds?

Well there's another reason to not use ChatGPT: "Tell me good slogans against the government" can get you arrested!