People will die because of this.

I'll try arguing in the opposite direction for the sake of it:

An "AI", if not specifically tweaked, is just a bullshit machine approximating reality same way human-produced bullshit does.

A human is a bullshit machine with an agenda.

Depending on the cost of decisions made, an "AI", if it's trained on properly vetted data and not tweaked for an agenda, may be better than a human.

If that cost is high enough, and so is the conflict of interest, a dice set might be better than a human.

There are positions where any decision except a few is acceptable, yet malicious humans regularly pick one of those few.

People will die because of this.

Yeah I'm going to make sure I don't take any new drugs for a few years. As it is I'm probably going to have to forgo vaccinations for a while because dipshit Kennedy has fucked with the vaccination board.

Yeah I'm going to make sure I don't take any new drugs for a few years. As it is I'm probably going to have to forgo vaccinations for a while because dipshit Kennedy has fucked with the vaccination board.

Text to avoid paywall

The Food and Drug Administration is planning to use artificial intelligence to “radically increase efficiency” in deciding whether to approve new drugs and devices, one of several top priorities laid out in an article published Tuesday in JAMA.

Another initiative involves a review of chemicals and other “concerning ingredients” that appear in U.S. food but not in the food of other developed nations. And officials want to speed up the final stages of making a drug or medical device approval decision to mere weeks, citing the success of Operation Warp Speed during the Covid pandemic when workers raced to curb a spiraling death count.

“The F.D.A. will be focused on delivering faster cures and meaningful treatments for patients, especially those with neglected and rare diseases, healthier food for children and common-sense approaches to rebuild the public trust,” Dr. Marty Makary, the agency commissioner, and Dr. Vinay Prasad, who leads the division that oversees vaccines and gene therapy, wrote in the JAMA article.

The agency plays a central role in pursuing the agenda of the U.S. health secretary, Robert F. Kennedy Jr., and it has already begun to press food makers to eliminate artificial food dyes. The new road map also underscores the Trump administration’s efforts to smooth the way for major industries with an array of efforts aimed at getting products to pharmacies and store shelves quickly.

Some aspects of the proposals outlined in JAMA were met with skepticism, particularly the idea that artificial intelligence is up to the task of shearing months or years from the painstaking work of examining applications that companies submit when seeking approval for a drug or high-risk medical device.

“I don’t want to be dismissive of speeding reviews at the F.D.A.,” said Stephen Holland, a lawyer who formerly advised the House Committee on Energy and Commerce on health care. “I think that there is great potential here, but I’m not seeing the beef yet.”

I have to quibble with you, because you used the term "AI" instead of actually specifying what technology would make sense.

As we have seen in the last 2 years, people who speak in general terms on this topic are almost always selling us snake oil. If they had a specific model or computer program that they thought was going to be useful because it fit a specific need in a certain way, they would have said that, but they didn't.

Or maybe that is part of the allure of automation: the eschewing of human responsibility, such that any bias in decision making appears benign (the computer deemed it so, no one’s at fault) and any errors - if at all recognized as such - become simply a matter of bug-fixing or model fine-tuning. The more inscrutable the model the better in that sense. The computer becomes an oracle and no one’s to blame for its divinations.

Your argument becomes idiotic once you understand the actual technology. The AI bullshit machine's agenda is "give nice answer" ("factual" is not an idea that has neural center in the AI brain), and "make reader happy". The human "bullshit" machine, has many agendas, but it would have not got so far if it was spouting just happy bullshit (but I guess America is a becoming a very special case).

Text to avoid paywall

The Food and Drug Administration is planning to use artificial intelligence to “radically increase efficiency” in deciding whether to approve new drugs and devices, one of several top priorities laid out in an article published Tuesday in JAMA.

Another initiative involves a review of chemicals and other “concerning ingredients” that appear in U.S. food but not in the food of other developed nations. And officials want to speed up the final stages of making a drug or medical device approval decision to mere weeks, citing the success of Operation Warp Speed during the Covid pandemic when workers raced to curb a spiraling death count.

“The F.D.A. will be focused on delivering faster cures and meaningful treatments for patients, especially those with neglected and rare diseases, healthier food for children and common-sense approaches to rebuild the public trust,” Dr. Marty Makary, the agency commissioner, and Dr. Vinay Prasad, who leads the division that oversees vaccines and gene therapy, wrote in the JAMA article.

The agency plays a central role in pursuing the agenda of the U.S. health secretary, Robert F. Kennedy Jr., and it has already begun to press food makers to eliminate artificial food dyes. The new road map also underscores the Trump administration’s efforts to smooth the way for major industries with an array of efforts aimed at getting products to pharmacies and store shelves quickly.

Some aspects of the proposals outlined in JAMA were met with skepticism, particularly the idea that artificial intelligence is up to the task of shearing months or years from the painstaking work of examining applications that companies submit when seeking approval for a drug or high-risk medical device.

“I don’t want to be dismissive of speeding reviews at the F.D.A.,” said Stephen Holland, a lawyer who formerly advised the House Committee on Energy and Commerce on health care. “I think that there is great potential here, but I’m not seeing the beef yet.”

It doesn't. I understand the actual technology. There are applications of human decision making where it's possibly better.

Or maybe that is part of the allure of automation: the eschewing of human responsibility, such that any bias in decision making appears benign (the computer deemed it so, no one’s at fault) and any errors - if at all recognized as such - become simply a matter of bug-fixing or model fine-tuning. The more inscrutable the model the better in that sense. The computer becomes an oracle and no one’s to blame for its divinations.

LLM does no decision making. At all. It spouts (as you say) bullshit. If there is enough training data for "Trump is divine", the LLM will predict that Trump is divine, with no second thought (no first thought either). And it's not even great to use as a language-based database.

Please don't even consider LLMs as "AI".

Oh good, a 60% chance you’ll get an ineffective or killer drug because they’ll use AI to analyze the usage and AI to report on it.

Even an RNG does decision-making.

I know what LLMs are, thank you very much!

If you wanted to even understand my initial point, you already would have.

Things have become really grim if people who can't read a small message are trying to teach me on fundamentals of LLMs.

Different types of AI, different training data, different expectations and outcomes. Generative AI is but one use case.

It's already been proven a useful tool in research, when directed and used correctly by an expert. It's a tool, to give to scientists to assist them, not replace them.

If you're goal to use AI to replace people, you've got a bad surprise coming.

If you're not equipping your people with the skills and tools of AI, your people will become obsolete in short time.

Learn AI and how to utilize it as a tool, you can train your own model on your own private data and locally interrogate the model to do unique analysis typically not possible in realtime. Learn the goods and bads of technology and let your ethics guide how you use it, but stop dismissing revolutionary technology because the earlier generative models weren't reinforced enough get fingers right.

I saw a paper a while back that argued that AI is being used as "moral crumple zones". For example, an AI used for health insurance acts allows for the company to reject medically necessary procedures without employees incurring as much moral injury as part of that (even low level customer service reps are likely to find comfort in being able to defer to the system.). It's an interesting concept that I've thought about a lot since I found it.

I wouldn't define flipping coins as decision making. Especially when it comes to blanket governmental policy that has the potential to kill (or severely disable) millions of people.

You seem to not want any people to teach you anything. And are somehow completely dejected at such perceived actions.

Oh good, a 60% chance you’ll get an ineffective or killer drug because they’ll use AI to analyze the usage and AI to report on it.