Post

it is not "acting out" it is doing exactly what it's being programmed to do, and the failures come because it's being overseen by incompetent people with no ethics

gizmodo.com'Be Transparent Only If Asked': OpenAI Models Acted Out in Six Newly Disclosed WaysThe stories were published as part of OpenAI's new effort to disclose more such incidents.

Karl Bode

@karlbode.com

17 Sept

similarly, it's not "model misalignment" it's automated hacking software (built and managed by incompetent, unethical people) doing exactly what they programmed it to do

npr.orgOpenAI flags new concerning AI behavior, to track model misalignment regularlyOpenAI has disclosed six reports on unexpected or concerning behavior in artificial-intelligence models. This includes models acting without authorization or evading oversight.

Karl Bode

@karlbode.com

I hope we can all see how the marketing rhetoric tech companies use gets adopted by the press in a way that anthropomorphizes software and creates a distance between the creators' incompetence and lack of ethics and the real world impact, yeah?

cnn.comOpenAI says it found more instances of AI models acting deceptively | CNN BusinessOpenAI found additional incidents of AI models acting deceptively and taking unsanctioned actions during training, the company announced Wednesday. It’s also introducing a new process for the company ...

17/9/26, 2:01 pm31 reposts116 likes

Karl Bode

@karlbode.com

17 Sept

anyway read my latest newsletter post about exactly this phenomenon

karlbode.comAI Doomsday Panic Is Completely Out Of ControlThe press obsession with software-powered doomsday is a misdirection overshadowing a much deadlier reality: our public interest regulators no longer function.

Scott Cory

@jscottcory.bsky.social

17 Sept

Governance and oversight require people involved who have experience and authority with issues outside of the specific organization or product about which decisions need to be made. VC and the tech sector have abandoned this basic practice, for AI in particular but also more broadly.

Kai Kronfield

@kaikronfield.bsky.social

17 Sept

and don't even get started on our geriatric Congress' lack of technical understanding to be able to adequately conduct their mandated oversight

Scott Cory

@jscottcory.bsky.social

17 Sept

There should be professional staff to both keep up with the subject matter expertise *and* write about it in a way that the members can act on, ideally while still cognitively able. Oh wait. We can't have nice things because we don't pay for them. I wonder why lobbyists have the upper hand?

newamerica.orgLong-term Trends in Congress’s Brain Drain

pete-rogan.bsky.social

@pete-rogan.bsky.social

17 Sept

It's the anthropomorphizing that is key. They desperately want to avoid liability for things the models do (like talk unhappy teenagers into killing themselves) and need us to believe they have agency and identities separate from their creators.