Trump-Appointed Federal Judge Scolds DOJ for Including Fake AI-Generated Case in Pleading

Say it aint so said:

Judge calls out Trump's DOJ over 'hallucinated' citation in ICE detainee case​

This is insane, Citing non existent case law:

View attachment 381599

I guess to keep up with the Trump HHS:

Trump officials downplay fake citations in high-profile report on children’s health​

So they are ignoring court orders, lying to judges, faking citations to justify their insane decisions, and now making up citations to bolster their legal briefings.
They have not one inkling of ethical behavior.

Click to expand...

Whoops. Well it's ok, lawyers who've done that oopsie before have merely just been disbarred a few times from what I remember.

As the use of AI increases, it's a hurdle that several different professionals will have to contend with in considering "immense volume with incredible speed" vs. "accuracy"

People will need to learn how to use it properly and know when to use advanced models, and when to use specialized "industry-specific" models.

The legal implications concern me less than some of the others (that profession was already full of opportunistic liars and schemers to begin with)

The potential medical implications concern me more.

In the legal world, there was a 90% chance that a defense lawyer was going to try to hide the truth or mislead the court anyway.

In the medical field, such an error means something goes undiagnosed, or gets misdiagnosed.
(although, I guess one could technically make a case that it happens already with humans in the medical field too)


The models are getting more sophisticated by the day, but people do need to keep in mind that there's a world of difference between the free stuff people can do through a web-browser, and the task/vocation-specific models that are focused on one thing.

It's an egg on the face moment of the administration if they've been leveraging the free-tier ChatGPT website for doing any serious research.

ThatRobGuy said:

As the use of AI increases, it's a hurdle that several different professionals will have to contend with in considering "immense volume with incredible speed" vs. "accuracy"

People will need to learn how to use it properly and know when to use advanced models, and when to use specialized "industry-specific" models.

The legal implications concern me less than some of the others (that profession was already full of opportunistic liars and schemers to begin with)

The potential medical implications concern me more.

In the legal world, there was a 90% chance that a defense lawyer was going to try to hide the truth or mislead the court anyway.

Click to expand...

The defense lawyers are also not allowed to cite fake cases even in defense of their client.

ThatRobGuy said:

In the medical field, such an error means something goes undiagnosed, or gets misdiagnosed.
(although, I guess one could technically make a case that it happens already with humans in the medical field too)


The models are getting more sophisticated by the day, but people do need to keep in mind that there's a world of difference between the free stuff people can do through a web-browser, and the task/vocation-specific models that are focused on one thing.

It's an egg on the face moment of the administration if they've been leveraging the free-tier ChatGPT website for doing any serious research.

Click to expand...

The US government didn't even check its citations. This administration has so flouted the norms that it is no longer granted the presumption of regularity.