HHS Asks All Employees To Start Using ChatGPT (404media.co) 64
An anonymous reader quotes a report from 404 Media: Employees at Robert F Kennedy Jr.'s Department of Health and Human Services received an email Tuesday morning with the subject line "AI Deployment," which told them that ChatGPT would be rolled out for all employees at the agency. The deployment is being overseen by Clark Minor, a former Palantir employee who's now Chief Information Officer at HHS. "Artificial intelligence is beginning to improve health care, business, and government," the email, sent by deputy secretary Jim O'Neill and seen by 404 Media, begins. "Our department is committed to supporting and encouraging this transformation. In many offices around the world, the growing administrative burden of extensive emails and meetings can distract even highly motivated people from getting things done. We should all be vigilant against barriers that could slow our progress toward making America healthy again."
"I'm excited to move us forward by making ChatGPT available to everyone in the Department effective immediately," it adds. "Some operating divisions, such as FDA and ACF [Administration for Children and Families], have already benefitted from specific deployments of large language models to enhance their work, and now the rest of us can join them. This tool can help us promote rigorous science, radical transparency, and robust good health. As Secretary Kennedy said, 'The AI revolution has arrived.'" [...] The email says that the rollout was being led by Minor, who worked at the surveillance company Palantir from 2013 through 2024. It states Minor has "taken precautions to ensure that your work with AI is carried out in a high-security environment," and that "you can input most internal data, including procurement sensitive data and routine non-sensitive personally identifiable information, with confidence."
It then goes on to say that "ChatGPT is currently not approved for disclosure of sensitive personally identifiable information (such as SSNs and bank account numbers), classified information, export-controlled data, or confidential commercial information subject to the Trade Secrets Act." The email does not distinguish what "non-sensitive personally identifiable information" is. HHS did not immediately respond to a request for comment from 404 Media. [...] The agency has also said it plans to roll out AI through HHS's Centers for Medicare and Medicaid Services that will determine whether patients are eligible to receive certain treatments. These types of systems have been shown to be biased when they've been tried, and result in fewer patients getting the care they need.
"I'm excited to move us forward by making ChatGPT available to everyone in the Department effective immediately," it adds. "Some operating divisions, such as FDA and ACF [Administration for Children and Families], have already benefitted from specific deployments of large language models to enhance their work, and now the rest of us can join them. This tool can help us promote rigorous science, radical transparency, and robust good health. As Secretary Kennedy said, 'The AI revolution has arrived.'" [...] The email says that the rollout was being led by Minor, who worked at the surveillance company Palantir from 2013 through 2024. It states Minor has "taken precautions to ensure that your work with AI is carried out in a high-security environment," and that "you can input most internal data, including procurement sensitive data and routine non-sensitive personally identifiable information, with confidence."
It then goes on to say that "ChatGPT is currently not approved for disclosure of sensitive personally identifiable information (such as SSNs and bank account numbers), classified information, export-controlled data, or confidential commercial information subject to the Trade Secrets Act." The email does not distinguish what "non-sensitive personally identifiable information" is. HHS did not immediately respond to a request for comment from 404 Media. [...] The agency has also said it plans to roll out AI through HHS's Centers for Medicare and Medicaid Services that will determine whether patients are eligible to receive certain treatments. These types of systems have been shown to be biased when they've been tried, and result in fewer patients getting the care they need.
Incoming (Score:2)
Re: (Score:1)
beat me to it
Nothing good... (Score:5, Insightful)
Re: (Score:3)
that ship sailed, crashed, burned and sank a long time ago
Re: (Score:2)
that ship sailed, crashed, burned and sank a long time ago
But the planks, oakum, and rigging keep getting replaced by the masses as it’s sinking and burning, so is it really the same ship?
Re: (Score:2)
that ship sailed, crashed, burned and sank a long time ago
It's more like that scene in the one pirates movie where Depp is riding the ship into port and steps off the mast onto the dock as it sinks, except our ship never fully sinks. We're just perpetually stuck in that moment right before he steps off, the impending doom feeling of that moment never fully manifesting, and never fully going away.
Re: (Score:1, Troll)
I dunno... at this point, it's not like ChatGPT would give *worse* advice than whatever you might get out of HHS. I would take medical advice from GPT before I would trust 45/7, RFK, or Oz to tell me that the sun is warm or water is wet. If you're taking advice from an AI, there's at least a chance... actually a fairly decent chance... you'll get the right answer. But even if it were only batting .300, that's still better than raving loons ranting on about how vaccines cause autism, ivermectin can cure
Re: (Score:3)
Re: (Score:1)
There is plenty of good that will come from this. RFKjr will get richer. His buddies will get richer. Citizens will become poorer, but, that is the purpose of citizens. To be raw food for the wealthy.
Something makes me feel uneasy... (Score:2)
Re:Something makes me feel uneasy... (Score:5, Insightful)
Too true, the article should have quoted "This tool can help us promote fake science, deceptive obfuscation, and systemic bad health."
South Park (Score:5, Funny)
It's like an episode of South Park... in real life.
What could possibly go right?
Re: (Score:2)
South Park holds a mirror up to society, only slightly distorted.
Re:South Park (Score:5, Insightful)
Except that today reality is so badly distorted that South Park isn't able to distort it more.
Don't worry (Score:5, Interesting)
Re: (Score:2)
You are assuming that they haven't negotiated some sort of separate, private ChatGPT instance built on training data which has been specifically tweaked to not allow that to happen (or else with an intermediate layer designed to intercept such suggestions before they reach the end user).
Re: (Score:2)
Re: Don't worry (Score:2)
I don't see where in this story government workers are 'required' to use AI, ChatGPT is being offered. Making a tool available is a lot less sinister than telling/requiring workers they must use the tool.
Re: (Score:2)
Highly accurate and reliable (Score:2)
Here is a list of the NFL teams that have six letters in their name: Bears, Bengals, Browns, Colts, Giants, Lions,Packers, Ravens, and Saints.
Re: (Score:2)
Since today was the Apple product announcement day, I thought it'd be interesting to ask ChatGPT if it knew what generation my Apple Watch was, based on the model number. I knew I could just Google it (and I think it says in the "about" section on the watch somewhere, but I didn't have the watch with me at the time), but I figured I'd give ChatGPT a shot. It responded with:
MU672LL/A is an Apple part/model number.
That specific one refers to the Apple USB SuperDrive – Apple’s slim external CD/DVD reader and burner that connects over USB-A
Apparently, my Apple Watch is so old it has turned into an external DVD drive. I hate it when that happens.
Re: (Score:2)
Re: (Score:2)
What model did you use? I used GPT-5 and it responded with Chiefs, Titans, Texans, Ravens, Giants, Eagles, Saints, and Browns. Looks right.
Whatever model, I just used the web browser version. You do point out an interesting quirk, sometimes the same prompt will yield different answers; not sure if that is a result of model differences.
Re: (Score:2)
Whatever model, I just used the web browser version. You do point out an interesting quirk, sometimes the same prompt will yield different answers; not sure if that is a result of model differences.
Nothing says reliable like non-deterministic output.
Re: (Score:2)
Whatever model, I just used the web browser version. You do point out an interesting quirk, sometimes the same prompt will yield different answers; not sure if that is a result of model differences.
FWIW, output is non-deterministic on purpose. Also, some models can be fed a seed that will then ensure the result is the (mostly the) same on repeated runs:
https://cookbook.openai.com/ex... [openai.com]
I wonder if there are some especially good seeds that often produce "better" results, somewhat like how Minecraft uses seeds and some of those produce some spectacular results, while most produce average results and some are awful.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
The question did not say "exactly six letters." All of those have at least six letters, if you include the city name. Don't forget about the Crunch Bird.
I would genuinely love to see (Score:5, Insightful)
I guess if you are just ignorant. And I don't mean ignorant as an insult I mean you just don't have a clue what's going on.
But I mean we have literally watched Trump illegally deploying troops, the trade war is spiking inflation and it's become clear he is all over the Epstein files in the worst way possible.
And at this point it's also clear Trump is the Republican party. He isn't just a figurehead or even just the head of the party he is the embodiment of everything that makes the Republican party the Republican party.
I know we will never get a truthful or straight answer from Republicans because the moment you do they would stop being Republicans but still...
That's the catch 22. If you have enough self-awareness to acknowledge everything above you can't stay Republican so you just have to bury it all so you can maintain that identity I guess.
Re: (Score:2)
Cause they hate black people and because they'll take lies from a strongman over hemming and hawing and PC crap from a black woman.
Re: (Score:2)
Re: (Score:2)
A distinction without a difference.
Re: (Score:2)
Re: (Score:2)
https://www.merriam-webster.co... [merriam-webster.com]
Re: (Score:2)
Do you really think that all people who voted for Trump, or even most of the people who voted for Trump, hate black people?
Do you really think that all people who self-identify as "Republican", or even most of the people who self-identify as "Republican", hate black people?
Re: I would genuinely love to see (Score:3, Interesting)
It likely depends on how you define "hate". But maybe the definition is not so important. Not everyone who supported Hitler in the thirties in Germany hated Jews, for some definition of "hate". The definition didn't make a difference back then either. The people who brought him to power and did not hate Jews are still responsible for the outcome of their support.
Re: (Score:2)
Winner winner chicken dinner!
Re: (Score:2)
The only definition that matters is how the person I responded to defines it.
Re: I would genuinely love to see (Score:2)
You are missing the point. The definition is irrelevant. The outcome matters. People who vote for Trump because he's an "economic genius" knowing that he will also oppress a certain minority, even if the oppression isn't what is motivating them, are responsible for the oppression of that minority. That's because actions matter. If you vote for a racist president then you are fostering racism and therefore you are racist; you telling yourself and others that it's not how you feel deep inside is irrelevant.
Re: (Score:2)
I am asking the poster what he thinks about hate. His definition of hate and how he used it all that matters.
I don't understand your post, at all. It makes no sense in the context of what I asked the other poster as to what he thinks about hate.
Re: I would genuinely love to see (Score:2)
You and the other poster might define "hate" differently and therefore disagree on a technicality. But the fact remains that people who voted for Trump are responsible for giving power to someone who wants to harm minorities. That's the important thing here. Focussing on the technicality is a distraction.
Re: (Score:2)
You and the other poster might define "hate" differently and therefore disagree on a technicality. But the fact remains that people who voted for Trump are responsible for giving power to someone who wants to harm minorities. That's the important thing here. Focussing on the technicality is a distraction.
What the fuck are you even talking about?
Let's review because it doesn't seem like we are on the same page.
Here is what the poster said:
Cause they hate black people and because they'll take lies from a strongman over hemming and hawing and PC crap from a black woman.
So my question was whether he really believed that. My question didn't have anything to do with the rest of the shit you are babbling on about. I simply wanted to know if his opinion was really what he stated it to be, or was it hyperbole or some other linguistic construction.
Surely it is that poster, not you or me, nor anyone else, who is the expert on that poster's opinio
Re: (Score:2)
The thing is, most people do not have a clue what is going in. The republicans just have even fewer of the rest.
Re: (Score:2)
The media environment is a rough one, podcasts, live streams, social media figures it's easy to stay awash in conservative messaging and the messaging is on point.
Look at any crisis and the admin pits out a message and allllllllll those channels, from Fox to Charlie Kirk to the Twitter accounts is hitting that message. Anything reported by left media can be discarded.
You can see the Epstein stuff start to poke holes in this though.
Re: (Score:2)
They're rich enough to benefit personally from Republican-backed tax cuts, and they don't care about what happens to anyone else.
Their religious beliefs tell them that they must elect a government that will ban gay marriage, and they don't care enough about what else happens.
Re: (Score:2)
Former Republican here. I voted Republican in every election until Trump. Even in 2016, I believed Trump was not conservative in any way, but rather, would say whatever his supporters wanted to hear. I believed he was about one and only one principle: Loyalty to Trump.
As a former Republican, I completely agree with your take. I don't get it, and I have many family members who are among those who still support Trump. I have become a bit of an outcast.
Use it? (Score:2)
Use it for what? Anything, everything?
Just go use it!
good grief
Re: (Score:3)
I came to post exactly thanks. Thanks!
I know of a $15B company who recently told one of its major divisions, "We need a program to start using AI".
No word as to what to use it for.
When asked, "What problem are you trying to solve?" there was no answer.
Re: (Score:2)
The problem it tries to solve is "too many jobs". This is both (1) well-understood in the business world, (2) not acceptable to say openly in many circles.
Probably the guy was unable to determine if the audience were more "workers" or "owners", and just froze up, not knowing which way to apply the spin.
I asked ChatGPT about banning vaccines for childre (Score:5, Insightful)
No, it is not a safe idea to ban vaccines for children. Here's why:
Vaccines Are a Proven Public Health Tool
Vaccines have been thoroughly tested and are monitored for safety and effectiveness. Childhood vaccines protect against dangerous and potentially deadly diseases like:
Measles
Polio
Whooping cough (pertussis)
Mumps
Rubella
Diphtheria
Tetanus
Meningitis
Hepatitis B
And more
ChatGPT, still smarter than RFK and the Florida republican party. https://www.bbc.com/news/artic... [bbc.com]
Re: (Score:2)
Too bad RFK will claim it’s biased and will switch to Grok.
Re: (Score:2)
Too bad RFK will claim it’s biased and will switch to Grok.
Even Grok says:
Reducing vaccine availability is generally a bad idea. Vaccines prevent millions of deaths annually—smallpox alone was eradicated, saving countless lives. The WHO estimates vaccines prevent about 6 million deaths yearly from diseases like measles, polio, and influenza. Lowering access increases risks of outbreaks, especially for vulnerable populations like children or the immunocompromised. For example, the 2019 measles resurgence in the U.S. was linked to vaccine hesitancy and reduced uptake.
Now, you can generally steer LLMs to produce a more biased output with appropriate prompting. "From the perspective of an anti-vaxxer, what are some of the reasons why one might oppose vaccine mandates?" will get you in the talking points ballpark. And if you want something more RFK-ish, just add "stark raving lunatic" to the prompt.
We come to to real reason at end of summary (Score:2)
"These types of systems have been shown to be biased when they've been tried, and result in fewer patients getting the care they need."
Just like other for profit health care systems have been developed to deny care. Our current regime is out to kill all poor people.
Re: (Score:2)
And keep redefining "poor"-status when the threat of running out of poor people looms?
Idiocy (Score:3)
RFK and most of the Trump administration don’t understand how LLMs work, they just assume it’s magically smarter than they are so they are content asking it questions and following it like an Oracle.
Nothing good will come of this. Once it tells RFK that ibuprofen causes autism and suggests putting antipsychotic meds into the water supply, he will follow it, because it’s “smarter” than he is.
Elon is angry (Score:2)
Elon Musk must be so pissed-off: This is what DoGE and xAI 'Grok' was meant to do. Instead, Palantir got all the pork.
I suspect he will ask CBP/ICE to investigate/kidnap OpenAI employees. It won't change anything but many rich people are sore losers and enjoy better people, wrongfully suffering.
It is a department in need of intelligence (Score:1)
Where did they dig up that old fossil? Was he Hoggle in Labyrinth [youtu.be]
Oh, damn, this smacks of religious fervor. (Score:4, Insightful)
Just the first paragraph of the summary, to prevent myself from going into mind-melting seizure territory.
Employees at Robert F Kennedy Jr.'s Department of Health and Human Services received an email Tuesday morning with the subject line "AI Deployment," which told them that ChatGPT would be rolled out for all employees at the agency. The deployment is being overseen by Clark Minor, a former Palantir employee who's now Chief Information Officer at HHS. "Artificial intelligence is beginning to improve health care, business, and government," the email, sent by deputy secretary Jim O'Neill and seen by 404 Media, begins.
This is literally marketing as religion. We have no data to back this up, we aren't actually waiting to see the effect it has, we're just gonna say it's improving things because "WE" desperately want to believe it improves things. We NEED it to improve things. Because we need that revenue generator, baby. This fervent "believe until we make it real" nonsense smacks of old-world religion, where they literally manifested things into existence simply by making the masses believe it. The problem is, the masses see this for the fraud it is, but these assholes are so up their own backsides they just keep preaching the same chorus, over and over and over and over, as they repeatedly thrust their AI everything deeply down our throats.
"Our department is committed to supporting and encouraging this transformation. In many offices around the world, the growing administrative burden of extensive emails and meetings can distract even highly motivated people from getting things done. We should all be vigilant against barriers that could slow our progress toward making America healthy again."
And there's the rub. We've spent several generations creating busywork nonsense that literally accomplishes jack-all, and prioritizing that busywork nonsense over actual work, which the management types have told us is vitally important to the bottom line, just to keep everybody busy. Now that we have something that sorta/kinda can automate some of that busywork and suddenly management types are all gung-ho about what a negative impact this busywork is on the average workday. Fuck you cunts. We've been saying that the entire god damned time. Meetings to support meetings with never-ending nonsense efail streams do nothing but waste time and keep people busy. Now we're automating the efail streams, making them longer and even more nonsensical in the process, and are automating summaries of those efail streams so that we don't negatively impact the humans involved in the process. Where's the benefit? Yes, you get to pay some other company a huge amount of money every year to make sure that the email streams through the servers are massive tomes of words, but if you're automating the writing and automating the summaries, WHERE. IS. THE. BENEFIT?
And no, profit for these AI companies is *NOT* a benefit for the end-user. We're shoehorning this shit into everything just to create more busywork, but now the busywork is on the computers instead of us. Um, yay? WTF?
Now a new point of blame... (Score:2)
Now a new point of blame...instead of the devil made me do it, it will be: "ChatGPT said so/told me to..." The new age beckons! (note the sardonic tone...)
JoshK.
non-sensitive personally identifiable information (Score:1)
This is an oxymoron.
Misleading headline (Score:1)
... not being told to use it - just that they are getting it.
Basically the license is being assigned to them.
RFK is trying to kill us all (Score:1)
RFK is trying to wipe out all of humanity.
Twice the work (Score:2)
At half the quality!