<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	xmlns:unherd="https://www.unherd.com"
	>

	<channel>
		<title>The mental elite - UnHerd</title>
		<atom:link href="https://unherd.com/tag/the-mental-elite/feed/" rel="self" type="application/rss+xml" />
		<link></link>
		<description>think again</description>
		<lastBuildDate>Thu, 19 Nov 2026 21:49:08 +0000</lastBuildDate>
		<language>en-US</language>
		<sy:updatePeriod>hourly</sy:updatePeriod>
		<sy:updateFrequency>12</sy:updateFrequency>
		<generator>https://wordpress.org/?v=7.1.5</generator>

<image>
	<url>https://unherd.com/wp-content/uploads/2019/09/cropped-favicon.jpg?w=32</url>
	<title>The mental elite - UnHerd</title>
	<link></link>
	<width>32</width>
	<height>32</height>
</image> 
<site xmlns="com-wordpress:feed-additions:1">241381703</site>				<item>
					<title>AI will create a new mental elite</title>
					<link>https://unherd.com/2026/11/ai-will-create-a-new-mental-elite/?edition=us</link>
											<comments>https://unherd.com/2026/11/ai-will-create-a-new-mental-elite/?edition=us#respond</comments>
										<unherd:featured_image>https://unherd.com/wp-content/uploads/2026/11/Harvard-old-time.jpeg?w=455</unherd:featured_image>
										<unherd:main_theme>Higher Education</unherd:main_theme>
					<dc:creator>
						<![CDATA[Richard Hanania]]>
					</dc:creator>
					<unherd:creator_img>https://unherd.com/wp-content/uploads/2025/12/image-7.jpg?w=96</unherd:creator_img>
																<unherd:article_ranking>3</unherd:article_ranking>
																<unherd:image_caption>
							<![CDATA[AI means that it’s no longer your grandfather’s Harvard. Credit: Getty]]>
						</unherd:image_caption>
						<unherd:twitter_bio>
							<![CDATA[<p>Richard Hanania<b data-stringify-type="bold">,</b> an <i data-stringify-type="italic">UnHerd </i>columnist, is the president of the Center for the Study of Partisanship and Ideology and blogs at <a class="c-link" href="https://www.richardhanania.com/" target="_blank" rel="noopener noreferrer" data-stringify-link="https://www.richardhanania.com/" data-sk="tooltip_parent">Substack</a>.</p>
]]>
						</unherd:twitter_bio>
						<unherd:twitter_handle>RichardHanania</unherd:twitter_handle>
															<unherd:author_bio>
						<![CDATA[Richard Hanania<b data-stringify-type="bold">,</b> an <i data-stringify-type="italic">UnHerd </i>columnist, is the president of the Center for the Study of Partisanship and Ideology and blogs at <a class="c-link" href="https://www.richardhanania.com/" target="_blank" rel="noopener noreferrer" data-stringify-link="https://www.richardhanania.com/" data-sk="tooltip_parent">Substack</a>.]]>
					</unherd:author_bio>
										<pubDate>Fri, 20 Nov 2026 00:00:15 +0000</pubDate>
					<updated>Fri, 20 Nov 2026 00:00:15 +0000</updated>
							<category><![CDATA[Uncategorized]]></category>
		<category><![CDATA[Artificial intelligence]]></category>
		<category><![CDATA[Cheating]]></category>
		<category><![CDATA[David Deming]]></category>
		<category><![CDATA[Higher education]]></category>
		<category><![CDATA[LLMs]]></category>
		<category><![CDATA[Professors]]></category>
		<category><![CDATA[The mental elite]]></category>
		<category><![CDATA[Universities]]></category>
					<guid isPermaLink="false">https://unherd.com/?p=1083208</guid>
																												<description>
								<![CDATA[And that’s a good thing]]>
							</description>
												
																									<content:encoded>
								<![CDATA[<p><span style="font-weight: 400">Artificial intelligence has triggered what can only be described as hysteria about the future of the education system. In the first month after ChatGPT was released,</span><i><span style="font-weight: 400"> The Washington Post</span></i> <a href="https://www.washingtonpost.com/education/2022/12/28/chatbot-cheating-ai-chatbotgpt-teachers/" target="_blank" rel="noopener"><span style="font-weight: 400">described</span></a><span style="font-weight: 400"> a state of “near-panic” among teachers and professors, and </span><i><span style="font-weight: 400">The Atlantic</span></i> <a href="https://www.theatlantic.com/technology/archive/2022/12/chatgpt-ai-writing-college-student-essays/672371/" target="_blank" rel="noopener"><span style="font-weight: 400">declared</span></a><span style="font-weight: 400"> that “The College Essay Is Dead.” </span><i><span style="font-weight: 400">New York</span></i><span style="font-weight: 400"> magazine </span><a href="https://nymag.com/intelligencer/article/openai-chatgpt-ai-cheating-education-college-students-school.html" target="_blank" rel="noopener"><span style="font-weight: 400">informed us</span></a><span style="font-weight: 400"> last year that “Everyone Is Cheating Their Way Through College.” The historian David Roach told </span><a href="https://www.newyorker.com/news/fault-lines/the-despair-of-the-professor-in-the-age-of-ai" target="_blank" rel="noopener"><i><span style="font-weight: 400">The New Yorker</span></i></a><span style="font-weight: 400"> that his “emotional response to all this is hard to describe, something between disgust and despair.”</span></p>
<p><span style="font-weight: 400">Essays have always been a staple of college classes. Today, AI can produce a passable paper on practically any topic in seconds. STEM fields and computer science are being hit just as hard, since large language models can do all the coursework in such fields, too. What happens to college when the metrics used to teach and gauge student achievement become worthless? </span></p>
<p><span style="font-weight: 400">It is true that it is easier than ever to cheat, and it seems to be highly likely that a large share of students are relying on LLMs in situations where they technically shouldn’t. The important question, though, is how much this matters, and whether what we are witnessing is the ruination of higher education — or merely a clearer view of the farce it had become long before the advent of generative models. </span></p>
<p><span style="font-weight: 400">The right answer, in short, is the latter: AI is merely revealing that most students were always ill-suited to higher education in the true sense, much of which was just about securing social status to begin with. AI’s wrecking of the old system, then, should be a boon to an exceptional elite: the few who were fit for </span><i><span style="font-weight: 400">higher education </span></i><span style="font-weight: 400">— the name itself is telling — in the first place.</span></p>
<p><span style="font-weight: 400">Let’s start by dispensing with a common misconception: that AI will inevitably make students dumber. It certainly can have that effect if use is unrestricted — but as you might expect given the capabilities of the technology, it can also provide ultra-competent tutoring on a mass scale. </span></p>
<div class="ref-ar"><a href="https://unherd.com/2026/04/how-ai-will-cure-populist-paranoia/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><a href="https://unherd.com/2026/04/how-ai-will-cure-populist-paranoia/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><img decoding="async" src="https://unherd.com/wp-content/uploads/2026/04/GettyImages-2204064825.jpg?w=300" /></a><div class="content"><a href="https://unherd.com/2026/04/how-ai-will-cure-populist-paranoia/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><span class="meta">More from this author</span></a><a href="https://unherd.com/2026/04/how-ai-will-cure-populist-paranoia/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><h5>Will AI kill off populism?</h5></a><p class="author"><a href="https://unherd.com/2026/04/how-ai-will-cure-populist-paranoia/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><span>By</span> Richard Hanania</a></p></div></a></div>
<p><span style="font-weight: 400">How the technology is used seems to determine its effects on student abilities. </span></p>
<p><span style="font-weight: 400">More important, the panickers completely miss the point of what education is for. Particularly at the higher levels, schooling has never been designed primarily to teach people how to think better, or even to impart useful information. Rather, degrees are a social signaling device: a way to tell others — most notably, employers — that the graduate was smart and conscientious enough to obtain a degree and can therefore be expected to follow similar bureaucratic processes in other realms. Wasting time and effort on social and bureaucratic signaling is bad for everyone. </span></p>
<p><span style="font-weight: 400">In recent days, David Deming, an economist and the dean of Harvard College, </span><a href="https://substack.com/home/post/p-225079575" target="_blank" rel="noopener"><span style="font-weight: 400">published a review</span></a><span style="font-weight: 400"> of the literature on the impact of AI on educational outcomes. This involved looking into more than 75 studies. Contrary to media reports, which tend to hype negative results, the research generally finds that AI helps with learning. It can cause problems when use is unrestricted, but practices like creating AI tutors designed to help with a class or using AI for feedback generally improve outcomes. </span></p>
<p><span style="font-weight: 400">This shouldn’t be surprising. As with practically every other technology, society is able to adjust. Moreover, how often do we worry about tutors being too smart, too knowledgeable, or too energetic? Since AI is off the charts in terms of some of the most important qualities we look for in teachers and professors, it would be strange if there were no way to harness the technology for good in the education sector. </span></p>
<p><span style="font-weight: 400">The larger point is that declines in learning truly don’t matter all that much, at least as far as the average student goes. Under a rather quaint view, education is about some combination of learning and becoming smarter. This is certainly true when it comes to basic skills like learning to read and doing arithmetic. It’s also true of the narrow slice of information a college student will pick up that will turn out to be directly relevant to his career. But the vast majority of what we learn in school is useless and quickly forgotten. We get degrees not primarily because doing so makes us smarter or because of what we learn, but because degrees are a way to signal to the world that we possess certain underlying traits: namely intelligence and conscientiousness.</span></p>
<p><span style="font-weight: 400">As the economist Bryan Caplan has argued, this explains features of education that we take for granted, but that would otherwise be mysterious. Why should we worry about cheating in the first place? People who actually want to learn a skill like dancing or cooking don’t cheat, because what they are learning is what matters to them. But most college students clearly don’t see the benefit to themselves of deeply comprehending what they are taught. One may think that they are simply being short-sighted, but the evidence indicates that they are basically correct. People </span><a href="https://psycnet.apa.org/record/1991-17502-001?doi=1" target="_blank" rel="noopener"><span style="font-weight: 400">generally do not retain</span></a><span style="font-weight: 400"> all that much of what they learn in school. </span></p>
<p><span style="font-weight: 400">Think also about how students react to news that class is canceled. They are practically always happy, which wouldn’t make sense if time spent in a classroom were bringing them value. Students behave as if they are in college to jump through hoops and get a credential that will help them in life, and their assumptions are, practically speaking, correct.</span></p>
<div class="su-pullquote su-pullquote-align-left">“Most college professors don’t want to think of themselves as DMV ladies. But that is what they are.”</div>
<p><span style="font-weight: 400">Education at the college level is therefore more like standing in line for a driver’s license than developing a skill. If we invented robots that could go to the local Department of Motor Vehicles and stand in line in place of humans, the same people would get their licenses in the same order, and nobody would be worse off. In fact, there would be a benefit, since people can find better things to do with their time. </span></p>
<p><span style="font-weight: 400">AI, then, simply makes it impossible to maintain the facade that a real and permanent impartation of skill and information is taking place in most higher-ed contexts. Most college professors don’t want to think of themselves as DMV ladies. But that is what they are. Society forces professors to spend a good portion of their working lives teaching bored undergrads about the subject they were passionate enough about to dedicate their lives to. But as the slouching posture of the undergrads attests, most others don’t care. </span></p>
<p><span style="font-weight: 400">How many fully accept the idea that what they are doing is ultimately useless? The students seem to be more willing to face the truth — hence, the widespread cheating that we see going on through the use of LLMs.</span></p>
<p><span style="font-weight: 400">Can the system survive once it is exposed as mostly fake? One possibility is just doubling down on the lie. This means trying to re-engineer college so that students are judged through assessments that bar LLM use. That would mean fewer take-home essays and problem sets counting toward final grades, with more emphasis on proctored exams and in-class essays. Blue books galore. We could eliminate the cheating, maintain the illusion that school is about learning, and go on as before.</span></p>
<p><span style="font-weight: 400">But that would be tantamount to deliberately gumming up the works at the local DMV. You know, just for the “rigor.” </span></p>
<p><span style="font-weight: 400">The second possibility is that professors simply accept cheating and don’t change all that much. Under this model, nobody is deceiving himself about what is going on. It is like the old Soviet joke: “We pretend to work, and they pretend to pay us.” Under this scenario, professors generate AI prompts as assignments, students use AI to do the work, and then professors grade the result with AI. Nothing changes about the system, except that everyone gets more time for sleeping, surfing TikTok, and socializing. Since things work out for all parties involved, nobody has an incentive to do anything differently.</span></p>
<p><span style="font-weight: 400">While it may be a good thing for people to work less hard and get the same results from schooling, this model will make it impossible to act as an honest person, go through college, and do as well as your classmates. Some students have a moral objection to lying and cheating, and it would perhaps be a terrible thing if we made such moral scruples into a major disadvantage in the education system. The signaling mechanism of education will have gone haywire. Schools (and thus employers) will be selecting for capable cheating as much as anything else. </span></p>
<p><span style="font-weight: 400">Which path seems more likely? Surveys from the last few years show that </span><a href="https://www.d2l.com/newsroom/new-research-reveals-ai-use-has-reached-a-tipping-point-in-higher-education/" target="_blank" rel="noopener"><span style="font-weight: 400">about</span></a> <a href="https://connect.chronicle.com/rs/931-EKA-218/images/GenAI-ResearchBrief.pdf" target="_blank" rel="noopener"><span style="font-weight: 400">half</span></a><span style="font-weight: 400"> of professors say that they have modified the types of assignments they gave in response to AI. That figure may seem high. But it is, in fact, shockingly low. Homework and take-home assignments are completely useless now, and yet </span><i><span style="font-weight: 400">half</span></i><span style="font-weight: 400"> of professors haven’t changed how they assess students?</span></p>
<p><span style="font-weight: 400">Moreover, even such findings might exaggerate the shift. A </span><a href="https://academic.oup.com/rcfs/article/15/4/1014/8746069" target="_blank" rel="noopener"><span style="font-weight: 400">2026 study</span></a><span style="font-weight: 400"> examining a major Israeli university found that 62% of professors said they had made changes in response to AI. But when the researchers looked at course syllabi, they found that “almost no courses have become less AI-compatible in the first two years after the release of ChatGPT.” Everyone is now cheating, and nobody cares.</span></p>
<div class="ref-ar"><a href="https://unherd.com/2026/08/have-you-corrected-your-racist-robot/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><a href="https://unherd.com/2026/08/have-you-corrected-your-racist-robot/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><img decoding="async" src="https://unherd.com/wp-content/uploads/2026/08/PNG-image-2.jpg?w=300" /></a><div class="content"><a href="https://unherd.com/2026/08/have-you-corrected-your-racist-robot/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><span class="meta">Suggested Reading</span></a><a href="https://unherd.com/2026/08/have-you-corrected-your-racist-robot/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><h5>Have you corrected your racist robot?</h5></a><p class="author"><a href="https://unherd.com/2026/08/have-you-corrected-your-racist-robot/?edition=us&#038;ref=refinnar" class="text-decoration-none border-0"><span>By</span> Sohrab Ahmari</a></p></div></a></div>
<p><span style="font-weight: 400">This puts the finding that AI can be beneficial in the classroom in a new light. Maybe it can, in fact, help students learn. But someone has to care enough to make sure that the technology is used in effective ways. Given the already-existing general indifference of both faculty and students, it is possible that nobody will. </span></p>
<p><span style="font-weight: 400">Rather than honesty becoming less valued as a general matter, cheating in school will likely come to be seen as less of a big deal. Nobody thinks, for example, that there is anything morally outrageous about driving five miles per hour over the speed limit, and a person can do so while being completely upstanding in other dimensions of her life. There is an understanding that the speed limit isn’t enforced </span><i><span style="font-weight: 400">that rigorously</span></i><span style="font-weight: 400">. Previously, cheating was seen as something you had to conspire to do. The friction involved ensured that most others also weren’t doing so on a mass scale. But now it is like speeding. And that means higher education, over time, will come to be seen as more of a sham, and something that even those with the toughest moral scruples don’t take seriously. </span></p>
<p><span style="font-weight: 400">This will, in a way, be a continuation of previous trends. Wokeness, the lowering of standards, and unserious academic degree programs have already made college seem largely a joke, with an understanding that getting credentialed and the quality of the university you attend are what really count.</span></p>
<p><span style="font-weight: 400">In the end, what truly matters about AI is not how it impacts the average student, but rather whether it helps the most talented students perform even better, and allows the most productive to eventually make breakthroughs in the fields they specialize in. There are early indications that this is indeed what is happening. </span></p>
<p><span style="font-weight: 400">The introduction of AlphaFold2, which is used in molecular-biology research, has </span><a href="https://www.ifo.de/DocDL/cesifo1_wp12462.pdf" target="_blank" rel="noopener"><span style="font-weight: 400">been associated</span></a><span style="font-weight: 400"> with an increase in the citation gap between the most and least productive scientists. One study published this year found that scientific fields that </span><a href="https://www.sciencedirect.com/science/article/pii/S0048733326000697?via%3Dihub" target="_blank" rel="noopener"><span style="font-weight: 400">incorporated</span></a><span style="font-weight: 400"> AI published more impactful work between 2005 and 2023. It is too early to talk about a general increase in scientific productivity, but either way, this is a metric to look out for. It barely matters how well most students are learning. </span><span style="font-weight: 400"> </span></p>
<p><span style="font-weight: 400">The highest-performing individuals are smart and highly motivated and, therefore, less in danger of seeing their cognitive skills atrophy. AI is probably another technology that — like the internet — is unquestionably good for the more talented and responsible among us, while posing some unintended dangers to others. Yet ultimately, performance at the highest levels of achievement is what matters for human progress. Due to AI, scientists are likely to be offering us many more astonishing technological developments and cures for diseases. </span></p>
<p><span style="font-weight: 400">Average students, meanwhile, may see some diminishment of their ability to learn and remember information they never use anyway, and get a lot more free time on their hands. This is a trade that is worth making.</span></p>
]]>
							</content:encoded>
																											<wfw:commentRss>https://unherd.com/2026/11/ai-will-create-a-new-mental-elite/?edition=us/feed/</wfw:commentRss>
															<post-id xmlns="com-wordpress:feed-additions:1">1083208</post-id>				</item>
						</channel>
</rss>
