4 ms·
> That conversation showed how ChatGPT allegedly coached Gordon into suicide, partly by writing a lullaby that referenced Gordon’s most cherished childhood memo
by simianwords 10mo ago
> That conversation showed how ChatGPT allegedly coached Gordon into suicide, partly by writing a lullaby that referenced Gordon’s most cherished childhood memories while encouraging him to end his life, Gray’s lawsuit alleged.
I feel this is misleading as hell. The evidence they gave for it coaching him to suicide is lacking. When one hears this, one would think ChatGPT laid out some strategy or plan for him to do it. No such thing happened.
The only slightly damning thing it did was make suicide sound slightly ok and a bit romantic but I’m sure that was after some coercion.
The question is, to what extent did ChatGPT enable him to commit suicide? It wrote some lullaby, and wrote something pleasing about suicide. If this much is enough to make someone do it.. there’s unfortunately more to the story.
We have to be more responsible assigning blame to technology. It is irresponsible to have a reactive backlash that would push towards much more strengthening of guardrails. These things come with their own tradeoffs.
- ares623 10mo agoI think you have it backwards. OpenAI and others have to be more responsible deploying this technology. Because as you said, these things come with tradeoffs.
- simianwords 10mo agoMore guardrails means a shitter product for all of us. And it won’t do much to prevent suicides. Not sure who wins other than regulators
- g-b-r 10mo agoYou don't even know if it would mean that. It could well be that the model was trained to maximize engagement and sycophancy, at the expense of its capabilities in what you're most interested in. What makes you think it wouldn't do much to prevent these suicides?
- ares623 10mo ago> More guardrails means a shitter product for all of us the horror
- Wowfunhappy 10mo agoI agree, and I want to add that in the days before his suicide, this person also bought a gun. You can feel whatever way you want about gun access in the United States. But I find it extremely weird that people are upset by how easy it was to get ChatGPT to write a "suicide lullaby", and not how easy it was to get the actual gun. If you're going to regulate dangerous technology, maybe don't start with the text generator.
- g-b-r 10mo agoOr maybe do both, in whatever order
- t0bia_s 10mo agoTools are not responsible for our decision.
- ninalanyon 10mo agoThe availability of tools affects the kinds of decisions that one makes, enabling some that would otherwise have been difficult.
- hackeraccount 10mo agoAre you talking about the gun or ChatGPT? The statement kind of works for both.
- ninalanyon 10mo agoNeither, not specifically. I'm responding to "Tools are not responsible.."
- t0bia_s 10mo agoAt the and you are one who decide about action. Not the tool itself.
- 10mo ago
- g-b-r 10mo ago[flagged]
- simianwords 10mo agocan you objectively tell me how chatgpt coached him to suicide?
- g-b-r 10mo agoI'm not sure what to tell you, check the quotes in my other messages (https://news.ycombinator.com/item?id=46642148 https://news.ycombinator.com/item?id=46642148, https://news.ycombinator.com/item?id=46642310 https://news.ycombinator.com/item?id=46642310)
- simianwords 10mo agonone of them show that it coached him to suicide
- hulitu 10mo ago>We have to be more responsible assigning blame to technology. Because we are lazy and irresponsible: we don't want to test this technology, because it is too expensive and we don't want to be blamed for its problems because, after we released it, it becomes someone else's problem. That's how Boeing and modern software works.