This is the mark of oppressive regimes, right? The details aren’t what matters (or, really, how they were drawn), but the plausible appearance of conclusions that fit the party line. Criticism? Just ignorable dissidents that no one important will hear.

We’re going to see a lot more of this, as chatbots are, unfortnately, rather good at making shallowly plausible walls of text. It’s easy for a lazy, incompetent person to do.

AI can be used to fact check papers too (for example, programatically following citations to see if they’re real, uncontroversial, or at least somewhat sensible), but it’s more technical to implement, and even if it wasn’t, that doesn’t even matter. This admin can simply shrug of any longwinded criticism as partisan and move onto the next controversy, knowing full well attention spans are too short to care. In other words, the information environment is the fundamental issue here.

Another investigation by The Washington Post found that at least 37 of the 522 citations appeared multiple times throughout the report. Notably, the URLs of several references included “oaicite,” a marker that OpenAI applies to responses provided by artificial intelligence models like ChatGPT, which strongly suggests its use to develop the report.

. . . [Bullshit Barbie] Leavitt described the errors as “formatting issues”

. . . The Washington Post notes that the MAHA report file was updated on Thursday to remove some of the oaicite markers and replace some of the non-existent sources with alternative citations.

And the article finishes off by letting the administration lie openly and unchallenged again.

Yeah. Most of us already know. But they’re going to have their way with us anyway, so what’s the use?

If you can’t get out, accept.

riddled with be