The reaction of the tech community to this story is extremely strange. If you type some sort of serious credible threat into a Saas platform, or post it on a message board or upload it as user generated content into any tech platform that is hosted in a “normal” country (by which I mean US, EU, Australia, UK, New Zealand, Canada and many other jurisdictions), your threat will very likely be reported to the police.

This isn’t a new phenomenon and it certainly isn’t an AI thing, it is a long-term policy posture that we as a society hold tech platforms accountable for serious threats that they are able to help to prevent that is implemented in a swathe of different legislation around the world, but it’s everywhere. This is clearly a privacy vs safety tradeoff and you can have all sorts of opinions about that, but this has been the case for a very long time (certainly since 2001).

Now different tech giants have different levels of competence at policing content on their own platforms but they all do it. As I say this is absolutely not anything to do with AI except to the extent that the AI companies are probably paying more attention to what is typed into their platform because they are using it to develop their models and the whole “AI safety” debate is evolving rapidly. But that’s merely a difference of degree.