DR. ZEYNEP TUFEKCI CLARIFIES A.I. 

AS CONFIRMED BY HER APPEARANCES AT THIS WEBSITE (see “Plato, The Internet, Education, and A.I.,” December 17, 2022; “Grok Goes Bonkers,” May 23, 2025; and “Deepfake—And Our Resulting Loss of Attention and Credibility,” August 14, 2025), I have immense respect for the insights of Dr. Zeynep Tufekci.

In her Guest Essay she addresses “The Unstoppable Force of A.I. Hype Is Meeting One Immovable Fact,” The New York Times, June 30, 2026, from which I glean tidbits that follow. Quoted items are Dr. Tufekci’s; other comments are mine.

The Hype. Dr. Tufekci posits, “So what would a fully automated future look like?” Then she cites Meta’s recent adoption of A.I. handling customer service: “Recognizing the opportunity that presented, scammers essentially talked the A.I. into turning over control of more than 20,000 Instagram accounts, including those of the Obama White House and a senior Trump administration official. Then the scammers lit up Telegram message boards with their delighted accounts of how easy it had all been.”

Note, the scammers didn’t just sit back and wait until A.I hallucinations screwed things up: they actively took control.

Other Screwups: “It was not a fluke. Air Canada disabled its chatbots after they mistakenly promised a customer a refund—and the customer sued and won. McDonald’s scuttled the bot taking orders at its drive-throughs after a number of viral videos showed it to be wildly dysfunctional. In one case, the bot mistakenly added hundreds of dollars of chicken nuggets to a customer’s order.”

Plausibility, not Reasoning. Dr. Tufekci explains that Large Language Models “can only assess which answers are probable, based on the data on which the models have been trained. And that holds true whether they’re trained on the full breadth of human output or only on peer-reviewed scientific articles.”

That is, inherent shortcomings reside in the process, not  just which sources are scarfed.

“So when an A.I. model follows a scammer’s carefully written prompts and gives away the keys to the kingdom—or when it responds to your earnest query with wild hallucinations—it’s not an aberration. It’s the technology working the way it was designed.”

The Exceptions. “The exceptions to that rule are jobs that occupy formal or verifiable domains. Coding is one such job. It relies on a structured, formal language that can be tested in real time.…The same goes for any other kind of work in which output is either verifiably right or wrong, functional or not functional, and can be definitively checked through an automated process.”

As a trivial example, “Is this product on the assembly line within its designed tolerances?”

But May I Speak to a Human, Please? “An overwhelming number of jobs, however, don’t work like that—not surgeon jobs and not customer service jobs and not fourth-grade teacher jobs. Those need the specialized technology of good old-fashioned human intelligence.”

I like Dr. Tufekci’s use of “good old-fashioned.” 

Generative A.I. and Its Problems. “Current A.I. is “made possible by vast amounts of data and computing power. They generate answers based not on truth or reasoning, but on probable connections among the data they have been fed. Hence the name: generative A.I.”

“We can’t fully control generative models. All we can do is train them up and then try to nudge them in the right direction. Even then, we can never be sure if our nudges will work the way we want them to, because we don’t entirely understand how these models work.”

Seeking At Least a Bit of Common Sense. “Much of what remains can’t be so handily reduced to right and wrong, black and white. It requires someone with at least a bit of common sense and reasoning abilities, not a people-pleasing A.I. chatbot that can be sweet-talked into doing things that defy logic.”

Dr. Tufekci’s Conclusion. “The sooner we update the way we think about the current state of A.I., the sooner we can all stop freaking out about the wrong things—and start preparing ourselves for the ways it really will transform our world.” 

One key, it would seem, is identifying whether “probably so” is an acceptable answer. Otherwise, as with many human endeavors, precise discernment and human intervention are required. Thanks, Dr. Tufekci, for prompting such cognition. ds  

© Dennis Simanaitis, SimanaitisSays.com, 2026