Will that nuance be understood by the models? It might be understood by people, of course, but will the models take into account that some people might support Palestine because of the alleged crimes against humanity commited by the attacker, and that some people might support Palestine because they are represented by HAMAS?

What will be the reasoning taken, what are the guidelines for AI to make this choice? If, as a government official, you are using AI for any action, you should be required to publish the model you are using and the specific prompt you are using.

An analysis could then be run on many different seeds for the model and prompt - some will label you as a 100% traitor to democracy, some will label you as a fighter for justice. That is almost certain.

Now what is the extent you are comforable with? Should we ban somebody whose actions are marked as traitorous in 10% of seeds? 20%? 80%? Do you actually believe anybody in the government, making these decisions, sees this nuance?

Normally if you criticise the israelian government in any way, the response is: "do not listen to this antisemite!"

How can you believe there's any good faith in this?

Yeah, the super magic big brother AI will definetely only target explicit supporters of Hamas and not anyone else, with no room for any abuse whatsoever :)

Well, consider me. I believed that leaving Saddam Hussein and the Baath party in charge of Iraq was preferable to finding some replacement, and I believe the same thing about Assad in Syria. Am I a Baathist?

Of course not. But is this view, that these Baathists should have remained in place support of them? In some way it must be, since it prefers them over something, right?

Even so, this is purely motivated by that I believe that their alternatives would slaughter or permit slaughter of all the minorities, in particular Christians, which is what happened at least in Iraq, so in practice it means that my view, rather than supporting any of these guys per se leads to opposition only to destabilizing these countries or other attempts to root them out.

So I suppose it depends on how you define support.

We don't know if and how the AI distinguishs between those two groups.

Yes, Trump during the campaign being careful to distinguish between protests against Israeli human rights violations in Palestine vs. those actually supporting Hamas during the campaign, the specificness of the searches the Administration has used when seeking to ban "DEI" topics from government funding, and the absence of hallucination in modern AI, all combine to give a high degree of confidence that this administration deploying AI to target Hamas supporters will not hit a variety of other (especially, but probably not exclusively, pro-Palestinian) targets.

This comment (while possibly well-intentioned) seems very disconnected from the ground level realities of how a large number of opinions are conflated (often deliberately) with support for Hamas: ranging from expressing horror at the killing and maiming of children, to expressing concern over arms exports to criticizing statements of specific politicians.

> This article talks about Hamas supporters, not people who support the freedom of Palestinian people.

Color me extremely skeptical.

Most people I have seen on the news advocating for Isreal to stop their genocide have been taxed of "supporting Hamas". I do not trust this administration, or "AI" to show any discernment.

You're quite naive if you haven't seen those two positions conflated.