I think about this finding a lot: "Students whose school uses AI for many reasons are more likely to report that they or a friend of theirs has had a back-and-forth conversation with AI, like chatbots, in potentially troubling ways." From @cdt.org's report:
Hand in Hand: Schools’ Embrace of AI Connected to Increased Risks to Students
Artificial intelligence (AI) has continued to alter the educational experiences of teachers, students, and parents during the 2024-25 school year. The frequency and variety of AI uses continues to gro...
cdt.org
Hypervisible @hypervisible.blacksky.app · 20h ago“…the researchers found that opening up to chatbots about personal issues made some people feel not better but worse.”
![AMERICAN ASSOCIATION OF UNIVERSITY PROFESSORS, LEONARD BRIGHT, VANITA REDDY, LANDON SADLER, and AARON GEORGE,
Plaintiffs,
V.
ROBERT L. ALBRITTON, JAY GRAHAM, DAVID
C. BAGGETT, JOHN W. BELLINGER, JAMES R.
BROOKS, MICHAEL A. HERNANDEZ III,
WILLIAM MAHOMES JR., KELLEY SULLIVAN GEORGIADES, and SAM TORN, in their official capacities as members of the Texas A&M University System Board of Regents,
Defendants.
CIVIL ACTION NO.
COMPLAINT FOR DECLARATORY AND INJUNCTIVE RELIEF
COMPLAINT
Plaintiffs American Association of University Professors ("AAUP"), on behalf of
its members in the Texas A&M University System ("Texas A&M"), and individually named member professors Leonard Bright, Vanita Reddy, Landon Sadler, and Aaron George (collectively, "Plaintiffs") challenge Texas A&M System Policy ("System Policy") 08.01's prohibition on teaching "gender and race ideology, or topics related to sexual orientation, and gender identity" in academic coursework (the "Censorship Policy" or the "Policy").
On its face, the Censorship Policy imposes a complete prohibition, in all core-
curriculum courses, on courses "advocat[ing]" particular disfavored views of race or gender, or
any topics related to sexual orientation or gender identity, but the Policy does not impose any corresponding ban on opposing views on these issues.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:jcats4toei4ybf5ibk3njty3/bafkreidml3fbnbiivajg7rverpx3nckfhv3zdron6klbxc7kaffoyslpty)
![A graphic poster reads "We have a deal! The ProPublica Guild has reached a deal with ProPublica's management on our first contract and has voted to ratify it. Swipe for the highlights [pointing emoji].](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:cdv2pqnwpgqvwp7i4h6qqm3m/bafkreicamcjhmrljmxkoelj7c44bud5joxgqjjqcwjjdq66mmnjljtqawe)
![D-CELL Games
Games: Unbeatable
Steam page
“Even if I’m just looking at this in the cold mercenary MBA freak way—the ‘I want to generate income as much as possible’ way—you can’t make a thing people want to buy if the thing you want to buy is the same thing everyone else has. And that’s what you fucking get with AI models. It’s definitionally part of how they work.
So all using [AI] will do is make you compete worse. You have to do laps on a hamster wheel to arrive at the opposite conclusion. Because the only way to actually make money on this shit besides a pump-and-dump is to be the guy selling the shovels. And even the shovels are starting to rust.
Even the stuff that sounds good or looks good for a given value of ‘good’ (i.e. it convincingly looks at a glance like it wasn’t generated with a machine, it’s ‘professional’) is only capable of spitting out inputs that existed before. You can feel it in the ‘choices’ it makes, because it doesn’t make choices; it chooses a value with a weight given by a model patched from every work at once on fast forward. The nutritional value of AI work is approximately identical to watching a 2000s horror movie trailer compilation at 4X speed while second-screening Maury transcriptions. It might be fun to look at, but it will kill you on the inside."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:5gdejlhcswjbzraadoltdpps/bafkreigiufrt3kp7efextkls2lz25nctkkjrso6ya4vbap7y72wkytlkem)
