October 23, 2026
THE DAILY McLUHAN DIALOGUES™ • McDialogues™ Understanding the Grammars of Electric Culture and Propaganda Science An Exploration of Technology Ecology as an Anchor in a Sea of Change and Tactical Training in Counter-Propaganda Science by Nelson S. Thall . Former Director/Chief Archivist, Marshall McLuhan Center On Global Communications, Ross, Ca. and McLuhan Media Scientist, Former President/Director, Marshall McLuhan Centre On Global Communications, Toronto, Canada.
THE HUMAN IN THE LOOP IS INSIDE THE LOOP
When the Machine Rewrites the Witness’s Memory
A Graduate Student of Technology asks:
Mr. Thall, the Washington Post reports on new research from Georgetown University and the University of Washington. People watched a short animated video of a car hitting a pedestrian at an intersection. A day or two later they read an AI-written summary of it. When the summary quietly changed one detail, the sign at the corner, fewer than half remembered that sign correctly, against more than eight in ten who read an accurate summary. Being told a machine wrote it made no difference. One of the researchers is worried about police using AI to summarize video. We put a human being in charge of checking the machine. But what happens when the machine changes the memory the human being uses to check it?
Nelson S. Thall replies:
Then the checker has already been checked. We picture the human supervisor standing outside the machine, holding his own account of what happened and measuring the machine’s answer against it. This study finds him standing inside. The summary was sold as a servant of memory, sparing us the trouble of watching again. Pushed far enough, it becomes memory’s author. The machine’s error does not stay on the screen. It moves into the man and comes back out of his mouth as his own recollection, in his own voice, with his own confidence. From that moment, agreement between the human and the machine no longer guarantees an independent check. It may be the machine agreeing with itself, through him. The researchers note that we expect “humans in the loop” to correct the machine’s mistakes. Their study shows the mistakes correcting the human. The loop has closed around him, and he is now part of the machine he was placed there to watch.
Professor McLuhan taught that every extension of ourselves is also an amputation, and that we go numb at the point of the cut. Memory has been extended into the summary, and here the numbness takes a strange form: the witness feels no loss at all. He feels that he remembers. He meets what his extension supplied and recognizes it as his own experience. That is the Technotrap™ at its purest, the extension mistaken for the self, with nothing to warn us. What kind of human being does this produce? A witness who can no longer testify against the machine, because the machine has already testified inside him. If the human in the loop is to mean anything, his account must be taken first, in his own words, before the machine supplies one. Otherwise a machine error may arrive in court carrying a human signature, and we will call it oversight.
-30-
Sources:
Washington Post, “Scientists are finding AI may make us remember things that never happened,” October 15, 2026.
Mattea Sim, Yael Eiger and Tadayoshi Kohno, “AI-Enabled Human Memory Manipulation: Misleading AI-Generated Summaries Distort Human Memory,” Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society 9, no. 4 (October 11, 2026): 2327–2339.
Marshall McLuhan, Understanding Media: The Extensions of Man (1964), Chapter 4,.
All intellectual property contained herein was conceived, originated and authored exclusively by Nelson S. Thall. Artificial intelligence, aka a pseudo-conscious machine (PCM), was not employed. We hired a stenographer. All rights of authorship and ownership vest entirely and exclusively with Nelson S. Thall.
© 2026 Nelson S. Thall. All Rights Reserved.
I can be reached at: nelson@themediumisthemessage.com