Here at ACS, we have analyzed metadata before. I think my favorite cases have been cell phone recordings of the original images. This is an example of some of the information that can be obtained.

Waiting for scientists figure out that in their pursuit of AI-themed funding, output from their models are not the greatest inputs for AI compared to actual data.

The violence in DC is so bad, our community suffered a terrible loss. If you visit Los Alamos, I think that's the idea being proposed here. "Man found guilty of murdering Spring mother when she dropped son off at Naval Academy in Maryland" abc13.com/post/michelle-…

If I had a student come in saying (s)he could teach a lab lesson based on AI, I would have the best time getting it all to fall apart.

Hope to read soon.. youtube.com/watch?v=ZlISx1…

This kind of fall (feet caught, head rotated downward) led to a fatality in one of my cases.

I have the highest respect for this author, and I agree with most of her points. But I think we also need to revisit the failure of Biosphere 2. Before we consider sending people to Mars, I suggest we understand what we're going to do if it fails. payloadspace.com/op-ed-nasas-ne…

The more I learn about social emotional learning (SEL), the more I picture this image.

While I'll bet that their Failure Mode and Effects Analysis included cooling system failure, this particular causal mechanism probably didn't.

The ejection is supposed to be facilitated for an emergency, which then makes it easier to accidentally trigger. The forces behind one will likely cause injury (near & long term), and that's balanced against the situation of losing the plane.

Academic freedom protects the pursuit of knowledge from partisan politics. No idea should be too dangerous for a public college classroom — and no politician should get to decide for students and faculty what they can and cannot discuss and debate. apnews.com/article/univer…

Safeguarding of women-wise, this appears to be a close call. The massive response to something that should be simple and obvious tells me there's a potential systemic safety failure mode to go looking for. memri.org/reports/al-azh…

I need to renumber my org safety failure red flags b/c of my manuscript rewrite, but this is the one where collecting data on the failure is blocked.