Does the marketing matter when the reason for the offending output is that the user spent significant deliberate effort in coaxing the LLM to output what it did? It still seems like MS Paint with extra steps to me.
I get not wanting LLMs to unprompted output “offensive content”. Just like it would be noteworthy if “Clear canvas” in MS Paint sometimes yielded a violent bloody photograph. But, that isn’t what is going on in OPs clickbait.
And, the thing is, LLMs are quite well protected. Look what I coaxed MS Paint to say with almost no effort! Don’t get me started on plain pen and paper! Which we put in the hands of TODDLERS!
I don’t get this. Why are so many countries willing to play Trump’s game? It seems a horrible long-term strategy to allow one country to hold global trade hostage this way. Shouldn’t we negotiate between ourselves, i.e., between the affected countries?
The idea should be: for us, exports of X, Y, and Z are taking a hit, and for you A, B, and C. So, let’s lower our tariffs in these respective areas to soften the blow to the affected industries. That way, we would partly make up for, say, lost exports to the US for cars, at the cost of additional competition on the domestic market for, say, soy beans; and vise-versa; evening out the effects as best we can.
With such agreements in place, we can return to Trump from a stronger position and say: we are willing to negotiate, but not under threat. We will do nothing until US tariffs are back to the levels before this started. But, at that point, we will be happy to discuss the issues you appear to see with trade inbalances and tariffs, so that we can find a mutual beneficial agreement going forward.
Something like this would send a message that would do far more good towards trade stability for the future.
This is my point. The case we are discussing now isn’t noteworthy, because someone doing it deliberately is equally “impressive” as writing out a disturbing sentence in MS Paint. One cannot create a useful “answer engine” without it being capable of producing something that looks weird/provoking/offensive when taken out of context; no more than one can create a useful drawing program that blocks out all offensive content. Nor is it a worthwhile goal.
The cases to care about are those where the LLM takes a perfectly reasonable conversation off the rails. Clickbait like the one in the OP is actually harmful in that they drown out such real cases, and is therefore deserving of ridicule.