A Clinical Trial Of AI
By
Leonard Zwelling
https://www.nytimes.com/2023/03/24/opinion/yuval-harari-ai-chatgpt.html?searchResultPosition=1
It is quite evident that Hollywood has bought into the notion that there are many realities, whether we are all in The Matrix or traveling in the multiverse of Everything, Everywhere, All At Once.
In a cogent and timely opinion piece in The New York Times on March 27, author Yuval Noah Harari (Sapiens) and his co-authors Tristan Harris and Aza Raskin warn of the dangers of AI. Is it a threat to humanity? Will it wipe us out? How could it do that?
Their hypothesis is that AI is taking over our language, the means by which we communicate with one another. Language is what they call, “the operating system of human culture.” If we are all plugged in, zooming with one another, and gleaning reality from what we read on social media, it is not a huge stretch to understand the potential dangers of AI. The computers could take over our language and thusly, our minds and our minds’ perception of reality.
The authors make the analogy that drug companies cannot roll out a new therapy without first testing the new treatment for safety and efficacy. What’s the rush with AI? The potential for AI to revolutionize medicine is vast (see NEJM 388:12 for several articles on AI in medicine), but are we willing to forego our own clinical judgment for that of a computer?
When I was an intern, I had a patient with a large mass in his upper right abdominal quadrant that everyone, including my very experienced attending physician thought was his enlarged liver. I didn’t. Now, on the Duke Medical Service, interns are the primary docs and I could order anything I wanted as long as it was on the proscribed list of tests. I ordered a liver scan. It was normal. We eventually biopsied his renal cancer to make the correct diagnosis. Undoubtedly, AI would have supported the judgement of my attending physician and my resident. It would have been wrong.
The authors point out that “social media was the first contact between AI and humanity, and humanity lost.” When the boomers get their news from Facebook, Instagram, and TikTok, we all may need Dr. Rick to prevent us from becoming our kids. These authors advocate for slowing the roll out of new AI tools until we can be sure that we can control them in the future. It is not hard to grasp that once we cannot tell whether what we are reading has been generated by another human or a machine, and this is going to happen if it hasn’t already, we can no longer trust what we perceive as real to really be real. I don’t know who’s calling you on the phone, but on our home land line it’s mostly computers.
As the authors write: “By 2028, the U.S. presidential race might no longer be run by humans.” Of course, that may provide an improvement to what we get in 2024.
Remember, we have already had a president that chose to fabricate reality and alter history. Donald Trump began his run for the White House with a lie about President Obama and where he was born and ended his presidency with a lie about who won in 2020. Once the political machinery is turned over to AI, goodness knows what “potential leader of the free world AI” will spew out to gain your vote. And AI could be the reason China wins if its man-made, Communist Party-generated version of the truth needs to compete with the United States’ computer-generated version of the truth. Just see “War Games” if you don’t believe that the next war will be fought by computers.
AI is powerful technology, more powerful than any drug, including the one that babysat me—television. Once we all become dependent on it for our very existence, our defense, our food supply, and our communications, it will be too late. Just because man thinks it up, does not make it good. We need to go slowly and see what AI can do for us as well as what AI can do to us.
Dr. Zwelling’s new novel, Conflict of Interest: Money Drives Medicine and People Die is available at:
barnesandnoble.com,
on amazon if you search using the title and subtitle, and
directly from the publisher Dorrance at: https://bookstore.dorrancepublishing.com/conflict-of-interest-money-drives-medicine-and-people-die-pb/
2 thoughts on “A Clinical Trial Of AI”
What a timely and cogent reflection on artificial intelligence, now abbreviated by capital AI, as though it is already a God.
It will be interesting to see how and if government, research institutes, and industry develop and evaluate the use of AI in medicine. It will have some important benefits, and the real questions in my mind relate to surgical decision-making and computer-assisted robotic surgery, a special interest of mine (vascular surgeon)
Great op-ed. Thanks again for keeping our minds fit!
Jeb: It may be time to pull the plug before it’s too late. Remember, we don’t really understand how AI really works.