"Terrorism" here would mean a major violent action: bombing datacenters rather than smashing individual waymos. Office shootings or assassinations would also count provided the correct ideologically motivation. This is not an exhaustive list.
I will not bet in this market.
Update 2026-04-27 (PST) (AI summary of creator comment): Regarding non-physical attacks:
The creator is disinclined to count cyber/non-physical attacks, though a sufficiently large-scale attack aimed at destruction of the system might qualify
Data theft (e.g. user data) generally won't count, as it typically has motivations other than anti-AI sentiment
Poisoning/subversion of AI systems is unlikely to count, as terrorism is considered overt by nature, while poisoning is typically covert and aims to subvert rather than disable
1,000@EspenJohannesen I feel desinclined from considering non-physical attacks, though a sufficiently large scale one might make to cut if aiming at destruction of the system.
Things like user theft, even at a large enough scale, will usually have different potential motivations other than anti-AI sentiment.
Poisoning... It's an interesting case. I'd say terrorism is by nature overt, and poisoning will in general only be effective if couvert... And it would in general not aim to disable the system, just subvert it