Post

If Elon Musk and half the Tesla C-Suite were currently in jail because Grok spent weeks generating CSAM-on-demand while they memed about it, Altman and the rest might be a bit more cautious about letting their little robots commit share price friendly press releasable crimes.

Tom Roberts@tpgroberts.bsky.social

I do think you basically charge a few of them with the cybercrimes they're committing and the "alignment" problem does suddenly start to disappear.

24/9/26, 4:39 pm107 reposts322 likes

jamesneill.bsky.social

@jamesneill.bsky.social

They don’t *let* them do anything. Check out Cal Newport’s podcast or his recent interview with Ed Zitron if you want to know what’s really going on.

Richard Walker

@phlebas.bsky.social

Indeed. These are programs which are doing what they were programmed to do. Not understanding what you've programmed is no more a defence than telling the judge "I didn't mean to kill him but was off my head on booze and amphetamines"

David Antoš 🇪🇺

@davidantos.cz

Your tweet is self-contradictory. In legal terms,you claim that every AI runaway model is a product of negligence. That is very unlikely to be true given how surprising and complex patterns were investigated and described. With the proliferation of open source models, this will become a big headache

Dmitry Grozoubinski

@explaintrade.com

If it was discovered that I ran a company that used a team of photoshoppers working remotely from North Korea to generate on-demand sexualized images of women without their consent, including minors, which I then posted live on Twitter, I do not think the response from authorities would be silence.

Mike Taylor 🦕

@mike.sauropods.win.ap.brid.gy

@explaintrade.com Who did this?

Dmitry Grozoubinski

@explaintrade.com

I'm drawing a parallel with Elon Musk's Grok X-bot generating underage CSAM...

Mike Taylor 🦕

@mike.sauropods.win.ap.brid.gy

@explaintrade.com Ooohhhh, I see. Anthopomorphizing the GAIs used to do this. Well, I detest Musk as much as anyone, but I do feel this is a bit of a stretch.

Third spruce tree on the left

@tezoatlipoca.mas.to.ap.brid.gy

@mike @explaintrade.com I think Dmitry is pointing out the duality of "its ok when the AI does it, but its not ok if a human does it. Why do we hold AI to different, more acceptable standards than humans, if the _harm_ caused is the same.. or worse?"

Mike Taylor 🦕

@mike.sauropods.win.ap.brid.gy

@tezoatlipoca BTW., is your username a riff on Monty Python's "How not to be seen" sketch?

Mike Taylor 🦕

@mike.sauropods.win.ap.brid.gy

@tezoatlipoca @explaintrade.com Yeah, I now get that that's what he was doing. My problem with this framing is that it makes the GAI itself equivalent to a human, which it obviously is not. The blame rests not with a poorly configured computer program, but with the people who configured it […]

sauropods.winOriginal post on sauropods.win

Chrononaut

@chrononaut.bsky.social

Hence one of the main selling points of AI bsky.app/profile/chro...

Chrononaut@chrononaut.bsky.social

"Nooo, it's the AI doing it, not me" This is one of the most attractive promises of AI - the ability to deflect blame to the technology. Whether we allow individuals and companies to get away with it will determine whether AI becomes a productive tool or follows in the footsteps of crypto.

sarfrin.bsky.social

@sarfrin.bsky.social

I think it's worse than crypto because it directly creates the awful shit rather than just being an easier way to launder the profits.

DigitalEraBeatnik

@jignpig.bsky.social

It has far more social influence, that's for sure. Hard agree. Shady money stuff has always gone on, AI is so much more than that.

Elthampug 🇬🇧🇺🇦

@elthampug.bsky.social

If X wasn’t American I suspect it would’ve been shut down like RT and Sputnik for disinformation alone!