Gebru had just described a medical scribe that recorded a patient as microdosing mushrooms, something she had never discussed, and that nobody caught. Her argument is that the existential-risk framing does active harm by feeding automation bias, rather than merely diverting attention from present-day failures. It is a direct counterweight to the warnings coming from the labs themselves this week.
…that we see today, whether it's cloud, whether it is ChatGPT. But when I wrote this paper, ChatGPT hadn't come out yet, but we saw the race to build larger and larger language models. And so that's the danger of building larger and larger language models that we were describing in this paper.
That they would essentially parrot people?
To parrot is to repeat back without understanding, right? So there was this whole, this whole existential risk narrative was happening back then too, if you can believe it. And so there was all this conversation about how OpenAI had claimed that GPT-2, the precursor to GPT-3, that powers chat GPT was too dangerous and too powerful to release. There was this conversation about whether GPTs can be ethical or they're creative and all this stuff. And so we really wanted to ground the conversation in the real issues. One of them, one of these issues is the environmental catastrophe, which a lot of people are now seeing, but we discussed it back then. And that was one of the main sections that Google people were unhappy with, the environmental cost. The other one is not documenting your data because you say you have too much data to document. The other one is deceiving people into believing that there is a mind behind the textual outputs that they're interacting with. And so there, that's where we really wanted to explain that these systems are parroting the patterns of their training data. And it's very dangerous when you're outputting text like that because when you have a plausible sounding test or very…
We're going to take a short break and then come right back.
AI tools can write code, access files, connect to the internet, and execute commands at a speed no human can match. That productivity is powerful, but it also means one bad instruction or compromise tool can quickly. Become a security incident. ThreatLocker helps organizations use AI without losing…
It's funny, I must be too much of an elder millennial because you say that there's too much automation trust. And I'm like, I'm so distrustful. I've got to call the bank and it's a robot. I'm like, nope, nope. I don't want to talk to it. I know. Yeah, exactly. So one of Anthropic's co-founders…