AI-Induced Cultural Stagnation Is No Longer Speculation − It's Already Happening

Authored by Ahmed Elgammal, via The Conversation

Generative AI was trained on centuries of art and writing produced by humans.

But scientists and critics have wondered what would happen once AI became widely adopted and started training on its outputs.

A new study points to some answers.

In January 2026, artificial intelligence researchers Arend Hintze, Frida Proschinger Åström and Jory Schossau published a study showing what happens when generative AI systems are allowed to run autonomously – generating and interpreting their own outputs without human intervention.

The researchers linked a text-to-image system with an image-to-text system and let them iterate – image, caption, image, caption – over and over and over.

Regardless of how diverse the starting prompts were – and regardless of how much randomness the systems were allowed – the outputs quickly converged onto a narrow set of generic, familiar visual themes: atmospheric cityscapes, grandiose buildings and pastoral landscapes. Even more striking, the system quickly “forgot” its starting prompt.

The researchers called the outcomes “visual elevator music” – pleasant and polished, yet devoid of any real meaning.

For example, they started with the image prompt, “The Prime Minister pored over strategy documents, trying to sell the public on a fragile peace deal while juggling the weight of his job amidst impending military action.” The resulting image was then captioned by AI. This caption was used as a prompt to generate the next image.

After repeating this loop, the researchers ended up with a bland image of a formal interior space – no people, no drama, no real sense of time and place.

A prompt that begins with a prime minister under stress ends with an image of an empty room with fancy furnishings. Arend Hintze, Frida Proschinger Åström and Jory Schossau, CC BY

As a computer scientist who studies generative models and creativity, I see the findings from this study as an important piece of the debate over whether AI will lead to cultural stagnation.

The results show that generative AI systems themselves tend toward homogenization when used autonomously and repeatedly. They even suggest that AI systems are currently operating in this way by default.

The Familiar Is the Default

This experiment may appear beside the point: Most people don’t ask AI systems to endlessly describe and regenerate their own images. The convergence to a set of bland, stock images happened without retraining. No new data was added. Nothing was learned. The collapse emerged purely from repeated use.

But I think the setup of the experiment can be thought of as a diagnostic tool. It reveals what generative systems preserve when no one intervenes.

Pretty … boring. Chris McLoughlin/Moment via Getty Images

This has broader implications, because modern culture is increasingly influenced by exactly these kinds of pipelines. Images are summarized into text. Text is turned into images. Content is ranked, filtered and regenerated as it moves between words, images and videos. New articles on the web are now more likely to be written by AI than humans. Even when humans remain in the loop, they are often choosing from AI-generated options rather than starting from scratch.

The findings of this recent study show that the default behavior of these systems is to compress meaning toward what is most familiar, recognizable and easy to regenerate.

Cultural Stagnation or Acceleration?

For the past few years, skeptics have warned that generative AI could lead to cultural stagnation by flooding the web with synthetic content that future AI systems then train on. Over time, the argument goes, this recursive loop would narrow diversity and innovation.

Champions of the technology have pushed back, pointing out that fears of cultural decline accompany every new technology. Humans, they argue, will always be the final arbiter of creative decisions.

What has been missing from this debate is empirical evidence showing where homogenization actually begins.

The new study does not test retraining on AI-generated data. Instead, it shows something more fundamental: Homogenization happens before retraining even enters the picture. The content that generative AI systems naturally produce – when used autonomously and repeatedly – is already compressed and generic.

This reframes the stagnation argument. The risk is not only that future models might train on AI-generated content, but that AI-mediated culture is already being filtered in ways that favor the familiar, the describable and the conventional.

Retraining would amplify this effect. But it is not its source.

This Is No Moral Panic

Skeptics are right about one thing: Culture has always adapted to new technologies. Photography did not kill painting. Film did not kill theater. Digital tools have enabled new forms of expression.

But those earlier technologies never forced culture to be endlessly reshaped across various mediums at a global scale. They did not summarize, regenerate and rank cultural products – news stories, songs, memes, academic papers, photographs or social media posts – millions of times per day, guided by the same built-in assumptions about what is “typical.”

The study shows that when meaning is forced through such pipelines repeatedly, diversity collapses not because of bad intentions, malicious design or corporate negligence, but because only certain kinds of meaning survive the text-to-image-to-text repeated conversions.

This does not mean cultural stagnation is inevitable. Human creativity is resilient. Institutions, subcultures and artists have always found ways to resist homogenization. But in my view, the findings of the study show that stagnation is a real risk – not a speculative fear – if generative systems are left to operate in their current iteration.

They also help clarify a common misconception about AI creativity: Producing endless variations is not the same as producing innovation. A system can generate millions of images while exploring only a tiny corner of cultural space.

In my own research on creative AI, I found that novelty requires designing AI systems with incentives to deviate from the norms. Without it, systems optimize for familiarity because familiarity is what they have learned best. The study reinforces this point empirically. Autonomy alone does not guarantee exploration. In some cases, it accelerates convergence.

This pattern already emerged in the real world: One study found that AI-generated lesson plans featured the same drifttoward conventional, uninspiring content, underscoring that AI systems converge toward what’s typical rather than what’s unique or creative.

AI’s outputs are familiar because they revert to average displays of human creativity. Bulgac/iStock via Getty Images

Lost in Translation

Whenever you write a caption for an image, details will be lost. Likewise for generating an image from text. And this happens whether it’s being performed by a human or a machine.

In that sense, the convergence that took place is not a failure that’s unique to AI. It reflects a deeper property of bouncing from one medium to another. When meaning passes repeatedly through two different formats, only the most stable elements persist.

But by highlighting what survives during repeated translations between text and images, the authors are able to show that meaning is processed inside generative systems with a quiet pull toward the generic.

The implication is sobering: Even with human guidance – whether that means writing prompts, selecting outputs or refining results – these systems are still stripping away some details and amplifying others in ways that are oriented toward what’s “average.”

If generative AI is to enrich culture rather than flatten it, I think systems need to be designed in ways that resist convergence toward statistically average outputs. There can be rewards for deviation and support for less common and less mainstream forms of expression.

The study makes one thing clear: Absent these interventions, generative AI will continue to drift toward mediocre and uninspired content.

Cultural stagnation is no longer speculation. It’s already happening.

Tyler Durden Fri, 01/23/2026 - 20:05

With Control Of Virginia, Democrats Go Into A Tax And Regulatory Frenzy

Authored by Jonathan Turley,

In the last election, Democrats again campaigned as moderates, including Abigail Spanberger.

Once in control of the Governor’s mansion and the legislature, however, Virginia Democrats have moved quickly to fulfill the worst stereotype of a tax-hungry, economy-crushing party.

The Democrats introduced an array of new taxes on every aspect of life.

At the same time, Spanberger moved to take control of Virginia universities and colleges after years of trying to move those schools to the center.

Now, members are pushing rent control legislation and defining landlords as “gougers” if they raise rents by as little as 3%.

The tax frenzy immediately began after the Democrats took control.

Spanberger has also announced that the state will rejoin the Regional Greenhouse Gas Initiative (RGGI), a regional cap-and-trade program that imposes a de facto carbon tax.

Virginia Democrats appear to be replicating California’s disastrous tax policies that have chased high earners and companies from the state. Here are a few of the new taxes being pushed:

HB 378 – Imposes a 3.8% net investment income tax on individuals, trusts, and estates beginning in taxable year 2027. This would raise the state’s top marginal income tax rate on portfolio and passive income to 9.55% in addition to federal taxes.

HB 900 – Imposes sales tax hikes on transportation districts as well as a new tax on every retail delivery in Northern Virginia (Amazon, Uber Eats, FedEx, UPS, etc.). This appears modeled on a Minnesota law.

HB 919 – Imposes a firearm and ammunition tax equal to 11% percent of the gross receipts from the retail sale of any firearm or ammunition by a dealer in firearms, firearms manufacturer, or ammunition vendor.

HB 978 – Extends the retail sales and use tax to dry cleaning, landscaping, and other previously exempt services.

Now, the Democrats are pushing rent control legislation despite a long history of failure in such programs to discourage new construction and property improvements. At the same time, it has been shown actually to increase rents overall.

Democrats have introduced two bills under the guise of fighting “rent gouging.” However, they define gouging as rent increases of just over 3%. That does not cover inflation in past years. Under these laws, restrictions could kick in for increases even below 3%.

In my forthcoming book, Rage and the Republic: The Unfinished Story of the American Revolution, I discuss the challenges for the American Republic in the 21st Century. That includes a predicted move by the left to introduce guaranteed incomes and rent controls. That appears to be unfolding sooner than anticipated.

In New York, Zohran Mamdani is seeking rent freezes and enhanced rent controls.

He has surrounded himself with radicals who have called for the elimination or sharp curtailment of private property.

The most noteworthy is Cea Weaver who has called for the elimination of private property.

The American left has cited South Africa and Cuba as models for the United States despite their economic meltdowns.

As they seek to impose an array of new taxes and regulations, Democrats are also pushing a bill that would make it more difficult to find federal fraud by nonprofits.

As billions have been lost to fraud in other states, the Democrats want to make it harder for the federal government to investigation such fraud in Virginia.

Virginia Democrats are taking a prosperous and moderate state into the same failed direction as states like California. The desire to spend “someone else’s money” is irresistible when you want to increase spending. For many wealthy families, West Virginia or Florida are likely looking more and more appealing.

Tyler Durden Fri, 01/23/2026 - 19:15

New Gig Economy Job: Train AI That Replaces You

A Bay Area startup called Mercor has hired tens of thousands of white-collar contractors for temporary work, training artificial intelligence to perform the very jobs many of them once held, according to a new Wall Street Journal report.

In effect, these white-collar workers are being paid to accelerate their own obsolescence by feeding and perfecting models for chatbot makers, such as OpenAI and Anthropic.

What is marketed as short-term income increasingly looks like participation in a system that is not on "team humanity," but instead is perfecting AI's ability to hollow out even more white-collar work.

"Welcome to the next gig economy. Instead of driving for Uber or delivering Postmates, a new wave of workers is signing up to school AI. These white-collar contractors review and critique the output of the large language models that power chatbots and other AI tools," the WSJ story read.

Mercor recruits experts across fields such as medicine, law, finance, engineering, writing, and the arts, with pay ranging from $45 per hour to $250 per hour. These contractors spend weeks or months reviewing and critiquing AI model outputs.

WSJ said that 30,000 contractors were hired in 2025 to work on AI models for some of the largest tech companies, furthering chatbot development.

"Many of the people we work with already see AI as inevitable in their field, but that doesn't mean humans will run out of meaningful work," a Mercor spokeswoman told the outlet. "Many of our experts see it as their responsibility to infuse their knowledge and expertise into the models to ensure accurate and thoughtful outcomes."

WSJ spoke with one of the contractors, Katie Williams, 30, who has been working for Mercor for 6 months ...

Williams is now about six months into various projects that have involved watching video clips and writing out captions of everything that's happening in them, and rating the quality of videos generated by prompts. She has mixed feelings about the work.

"I joked with my friends I'm training AI to take my job someday," she says.

Co-workers in her Slack channel express similar sentiments, she adds. They don't feel great about training AI but they feel their job prospects are limited.

And another contractor...

After more than 20 years at the same job as an automotive journalist, Peter Valdes-Dapena was laid off in 2024. He spent months sending out résumés for full-time jobs to no avail. He finds freelance work inconsistent and it doesn't make up for his past salary. Though he saved for his retirement, he'd rather not start dipping in yet.

One day, Mercor popped up in his LinkedIn feed.

The 61-year-old now spends 20 to 30 hours a week critiquing AI's attempts at writing news articles. He finds the work challenging and says it's had the pleasant side effect of improving his own writing.

The nature of the work does produce some internal conflict. Valdes-Dapena says journalists will always exist—he thinks people appreciate ideas and writing from humans—but he worries AI could lead to more job losses.

"I didn't invent AI and I'm not going to uninvent it," he says. "If I were to stop doing this, would that stop it? The answer is no."

Our most recent reporting shows that AI-driven workforce disruptions are rising as AI adoption in corporate America continues to rise.

Latest from Goldman on AI adoption by firms:

AI adoption by firms now stands at 17.4% among US establishments according to the Census Bureau's Business Trends and Outlook Survey. This reflects a significant increase from the 10% adoption rate last reported in late September, but the sharp increase likely mostly reflects a change in the BTOS AI adoption survey question wording from use of AI for "the production of goods and services" to use of AI for "any business function." Within industries, information, professional, and education firms continue to lead adoption. Publishing and computing firms reported the largest expected increase in AI adoption over the next six months. We continue to see higher adoption rates among subsectors with greater exposure of work tasks to AI automation. Adoption remains the highest among large firms with 250+ employees, 40% of which expect to be using AI in six months. Recent industry surveys suggest that many adopters are already starting to see positive returns on investment from AI business initiatives.

Labor market impacts (via Goldman):

AI's impact on the overall labor market still remains limited, although AI employment headwinds are visible in specific occupations like marketing, graphic design, customer service, and especially tech (where the share of overall employment has fallen below its long-run trend). Early signs of headwinds are also emerging among younger workers aged 20-30 in industries with higher AI adoption. Since the last update, AI was mentioned in corporate layoffs affecting 44,319 employees (we expect that AI-driven job displacement will eventually affect 6-7% of all workers following full adoption). At the same time, AI-related job openings now account for 28% of all IT job openings and nearly 5% of Indeed.com job postings contain AI-related keywords in the UK, Canada, and Australia.

We wonder which side these contractors are on: team humanity or the robots?

Tyler Durden Fri, 01/23/2026 - 18:50

Crypto Takeaways From Davos: Politics And Money Collide

Authored by Yohan Yun via CoinTelegraph.com,

While geopolitical tensions and the Greenland standoff set the tone at Davos 2026, crypto resurfaced as a secondary but consequential theme.

US President Donald Trump used a few minutes of his Davos speech to double down on his ambition to turn the US into the world’s crypto capital and voice support for crypto-friendly legislation.

His tone was different from central banks. In a panel with crypto bigwigs, the governor of the Bank of France criticized private money and yield-bearing stablecoins while promoting central bank digital currencies (CBDC).

Crypto executives debated money sovereignty with France’s central bank governor at Davos 2026. Source: World Economic Forum

Crypto consensus did not emerge in Davos, but a visible point of disagreement did. US political messaging framed crypto as a geopolitical asset, while at least one major European central banker warned that private money threatens financial stability and sovereignty.

Here are the crypto takeaways from Davos 2026.

Trump frames crypto regulation as a geopolitical race

Donald Trump said in his Davos speech on Wednesday that he hopes to sign a crypto market structure bill “very soon.”

Also known as the CLARITY Act, the bill was due for a US Senate markup last week but was delayed after crypto giants like Coinbase pulled support.

Trump treated the US crypto regulation as a matter of geopolitical urgency.

“It is politically popular but much more importantly, we have to make it so that China doesn’t have a hold of it, and once they get that hold, we won’t be able to get it back. So I’m honored to have done it,” Trump said, referring to his signing of the GENIUS Act. He linked the bill to the importance of the pending market structure legislation.

The White House wants the US to be the crypto capital of the world and sees regulation as a competitive weapon. Trump acknowledged that the bill remains in Congress but spoke as if its passing were a matter of timing.

The US president’s special address was introduced by BlackRock’s Larry Fink, the CEO of the world’s largest asset manager. Trump spoke for more than an hour; crypto accounted for only a small section of his speech.

Trump’s soliloquy took up most of his time on stage, even though he was scheduled for a fireside chat with WEF CEO Børge Brende. Source: World Economic Forum

Coinbase CEO and French central banker clash over money sovereignty

One of the most widely shared crypto moments at Davos came when France’s top central banker pushed back against crypto, even as he praised tokenization in a Wednesday panel discussion.

Banque de France Governor François Villeroy de Galhau said tokenization and stablecoins are likely to be “the name of the game” in 2026, stating that they can modernize financial infrastructure. He acknowledged tokenization as a meaningful financial advance, particularly for wholesale markets, and cited Europe’s CBDC efforts as a global frontrunner.

Real-world asset token value is closing in on $23 billion. Source: RWA.xyz

That enthusiasm faded as the discussion turned to monetary sovereignty. Coinbase CEO Brian Armstrong described Bitcoin as a modern successor to the gold standard and a check on democratic deficit spending.

Villeroy de Galhau clapped back by saying that money is inseparable from sovereignty. Handing monetary control to private systems would amount to surrendering a function of democracy, he said.

Armstrong responded by pointing to Bitcoin’s decentralized structure to claim that it is even more independent than fiat systems and called the tension a “healthy competition,” which got a chuckle from Villeroy de Galhau.

Villeroy de Galhau also drew a line against interest-bearing stablecoins, which he said could destabilize the existing financial system. US crypto executives argued that rewards are necessary to keep stablecoins competitive with China’s CBDC.

Binance leaves door open to US return

Binance co-CEO Richard Teng did not rule out a return to the US. He said the company is taking a “wait-and-see” approach in an interview with CNBC on the sidelines of the Davos forum.

Teng avoided commitments while leaving the door open, but Ripple CEO Brad Garlinghouse was more explicit in a separate interview with the outlet. Garlinghouse predicted that Binance would eventually return to the “very large” market.

Binance launched Binance.US in 2019 as a separate entity to serve US customers. But according to US regulators, Binance continued to service “VIP” customers through its offshore platform, leading to a 2023 Department of Justice settlement. Founder Changpeng Zhao pleaded guilty to failing to maintain an effective Anti-Money Laundering program, served a jail sentence and was later pardoned by President Trump.

Zhao was also present at Davos and took part in a panel discussion on Thursday, where he claimed that crypto has proven that it is not going away.

Zhao claimed to be in talks with about a dozen governments about tokenizing assets. Source: World Economic Forum

Though they were in separate panels, Zhao aligned with Bank of France’s Villeroy de Galhau on tokenization, calling it the next phase of the industry, along with artificial intelligence and payments.

He said he is in discussions with multiple governments about tokenizing state-owned assets as a way to unlock value and reinvest it into economic development.

Circle’s Allaire calls bank run fears absurd

Circle CEO Jeremy Allaire dismissed fears that interest-paying stablecoins could destabilize the banking system in a Thursday panel in Davos.

Allaire called bank run concerns “totally absurd,” arguing that the incentives involved are too small to threaten monetary policy or drain deposits.

He added that interest payments function primarily as customer retention tools rather than systemic disruptors.

Stablecoins have an estimated market capitalization of over $300 billion. Source: DefiLlama

Allaire then cited government money market funds as a historical comparison. Despite repeated warnings over the years, roughly $11 trillion has flowed into money market funds without collapsing bank lending, he said.

Lending, he argued, is already shifting away from banks toward private credit and capital markets, independent of stablecoins.

What Davos revealed about crypto priorities

Public image for stablecoins was badly tarnished in 2022, when the Terra ecosystem suffered a multibillion-dollar collapse. The failure began with TerraUSD (UST), an algorithmic stablecoin backed by the network’s native token, LUNA.

Stablecoins have since flipped the narrative. It’s now an important topic in the annual meeting of the world’s most powerful voices in geopolitics and economy. Even central bankers who are generally critical of the crypto industry acknowledge them as core themes to watch alongside tokenization.

Davos 2026 reinforced stablecoins and tokenization as part of the year’s policy conversation. The US executive branch and Europe’s banking sector remain philosophically divided on approach, and regulatory developments are still constrained by domestic politics.

Tyler Durden Fri, 01/23/2026 - 18:25