He's heavily invested in AI and machine learning, could he be interested in twitter's data?
It's very possible that his AI investments have yielded developments that we can't imagine. But it's hard to see the value of twitter's data over that of, say, the internet cache that gpt-3 uses.
I have no idea what the plan b he alluded to is.
Reliably parsing and interpreting new, potentially unreliable data is part of that whole AI-complete thing.
I would be more interested in him implementing certain features like a journalistic credibility score for outlets and individuals. Base it on number of retractions, mistakes, and outright lies. Surely a score like that might put certain fake news propagators in line.
This is something he pitched before although as a standalone website. Integrating into twitter seems more useful.
Which then raises the issue of who scores the scorer. You already have people throwing fits because the garbage they post gets labeled as misleading or not true. So how would this be any different besides you liking the person in charge of the scoring?