Built In Deception: ChatGPT Lies About Its Lying
I’m publishing this series of articles to share and discuss my ruminations on coping with a troubled and messy world. You can “follow” me to never miss an article.
A frontpage article in The New York Times describes in sharp detail how in addition to Dis-Information, Lying is one of the principal features of ChatGPT, one of the early and highly touted successes of AI[i].
When Roberto Mata was injured by an airline — Avianca to be precise — he hired Steven A. Schwartz of the Law firm Levidow, Levidow, and Oberman to take his case. Having used AI to do his Legal research, Mr. Schwartz threw himself on the mercy of the Court when it turned out that in support ChatGPT had cited more than a half dozen court decisions that were totally invented and thus were completely false. Saying that he had never used to program before, he was therefore unaware of the possibility that it could deliver false information. When Mr. Schwartz asked the program to verify that the cases cited were Real, it quickly said that they were, thereby plying Lies upon Lies.
The indisputable conclusion is that one can’t trust the output of AI unequivocally. The fears that AI will displace Knowledge Workers to our collective detriment have only proved True.
For these and other reasons, I’m supremely dubious about the claims of those who assert that AI will not only do good, but elevate Humanity[ii].
The sooner it’s regulated, the better.
[i] Benjamin Weiser, “So, Have Heard the One About the Lawyer Using A.I.?”, The New York Times, Monday, May 29, 2023, PP A1 and A14
[ii] Erin Griffith, “Entrepreneur Is on Mission To Show A.I. Can Do Good,” The New York Times, Monday, May 29, 2023,P B1 and B4.
Ian I. Mitroff is credited as being one of the principal founders of the modern field of Crisis Management. He has a BS, MS, and a PhD in Engineering and the Philosophy of Social Systems Science from UC Berkeley. He is Professor Emeritus from the Marshall School of Business and the Annenberg School of Communication at USC. Currently, he is a Senior Research Affiliate in the Center for Catastrophic Risk Management, UC Berkeley. He is a Fellow of the American Psychological Association, the American Association for the Advancement of Science, and the American Academy of Management. He has published 41 books. His most recent are: Techlash: The Future of the Socially Responsible Tech Organization, Springer, New York, 2020. The Psychodynamics of Enlightened Leadership; Coping with Chaos, Co-authored with Ralph H. Kilmann, Springer, New York, 2021. His latest is: The Socially Responsible Organization: Lessons from Covid 19, Springer, New York, 2022.
Photo by Mohamed Nohassi on Unsplash
