Credible accusations must be made or linked to in the comments to count for resolution. Once a credible accusation is posted, the question will resolve a month later, to give time for other accusations to surface.

Deaths caused through any causal mechanism count, as long as they aren't too butterfly-effect-y. They must be specific deaths traced to the model. Feel free to ask about specific scenarios and I'll answer whether they count.

Rulings:
10-3-2023: If there are multiple listed organizations responsible for training the model (e.g. one organization did pretraining and the other did fine-tuning), it resolves to the organization that spent the most compute directly on training (for current models, this only counts the compute costs of the forward and backward passes).

10-3-2023: The "Facebook" answer should be interpreted to refer to the the whole Meta corporation, the "Meta" answer will not be chosen.

11-7-2023: Since there's a chance that resolution of this question comes down to a judgement call, I won't bet on this market.