When anthropology researcher Ashley McDermott was doing fieldwork in Kyrgyzstan a few years ago, she says many people voiced the same concern: Children were losing touch with their indigenous language. The Central Asian country of 7 million people was under Russian control for a century until 1991, but Kyrgyz (pronounced kur-giz) survived and remains widely spoken among adults.

McDermott, a doctoral student at the University of Michigan, says she also heard that some kids in rural villages where Kyrgyz dominated had spontaneously learned to speak Russian. The adults largely blamed a singular force: YouTube.

McDermott and a team of five researchers across four universities in the US and Kyrgyzstan have released new research they believe proves the fears about YouTube’s influence are valid. The group simulated user behavior on YouTube and collected nearly 11,000 unique search results and video recommendations.

What they found is that Kyrgyz-language searches for popular kid interests such as cartoons, fairy tales, and mermaids often did not yield content in Kyrgyz. Even after watching 10 children’s videos featuring Kyrgyz speech to demonstrate a strong desire for it, the simulated users received fewer Kyrgyz-language recommendations for what to watch next than, surprisingly, bots showing no language preference at all. The findings show YouTube prioritizes Russian-language content over Kyrgyz-language videos, especially when searching or browsing children’s topics, according to the researchers.

“Kyrgyz children are algorithmically constructed as audiences for Russian content,” Nel Escher, a coauthor who is a postdoctoral scholar at UC Berkeley, said during a presentation at the school last week. “There is no good way to be a Kyrgyz-speaking kid on YouTube.”

McDermott recalls one frustrated Kyrgyzstani mother in 2023 explaining that she paid the internet bill a day late each month to regularly have one day without internet and, thus, YouTube at home.

YouTube, which has “committed to amplifying indigenous voices,” did not respond to WIRED’s requests for comment. The researchers are attempting to meet with YouTube’s parental controls team to discuss the potential for language filters, according to Escher.

The researchers say their work is the latest to show how online platforms can reinforce colonial culture and influence offline behavior. Under Soviet control, people in Kyrgyzstan had to learn Russian to succeed. Today, many adults are fluent in both Russian and Kyrgyz, with Russian remaining important for commerce. Kids are required to learn at least some Kyrgyz in school. But many spend several hours a day online, and watching YouTube is the leading activity, McDermott says. Quoting from Russian language videos is common, whether creators’ refrains like “Let’s do a challenge,” adaptations of American words such as “cringe,” or parroting accents and syntax.

In one of the researchers’ experiments, they searched for several subjects which are spelled the same in Russian and Kyrgyz, including Harry Potter and Minecraft. The results were predominantly Russian. Overall, just 2.7 percent of the videos the research team analyzed appeared to even include ethnically Kyrgyz people.

YouTube “socializes youth to view Russian as the default language of entertainment and technology and to view Kyrgyz as uninteresting,” the researchers wrote in a self-published paper accepted to a social computing conference scheduled for October.

The researchers say there is ample Kyrgyz-language children’s content for YouTube to promote. In 2024, the 35th-most viewed channel on YouTube across the world was D Billions, a Kyrgyzstan-based children-focused content studio with a dedicated Kyrgyz-language channel that has nearly 1 million subscribers.

The annual showcase at the Centre for Emerging Technology and Security (CETaS) kicked off with a discussion on the implications of Claude Mythos. 

Opening the conference, Alexander (Sacha) Babuta, director of CETaS at the Alan Turing Institute, said that Anthropic’s latest frontier model, Claude Mythos Preview, demonstrates major improvements in mathematics, cyber security, software engineering and automated vulnerability detection.

While the model can identify and autonomously exploit previously undiscovered vulnerabilities in real-world systems, he described an optimistic outlook of how Claude Mythos Preview could be used to secure enterprise IT. “Companies can use models like Anthropic Mythos to rapidly discover vulnerabilities in their own systems and patch them to strengthen digital security for everyone,” said Babuta. 

A study of the cyber crime community between the release of ChatGPT in 2022 and the end of 2025 revealed that cyber crime forums played host to a number of “dark AI” products.

These are claimed by their owners to be homegrown or extensively retrained and jailbroken large language models (LLMs) customised and tailored for cyber crime. But despite generating some early enthusiasm on the forums, these have made little impact to date, Ben Collier, senior lecturer at the University of Edinburgh, said in a presentation discussing the findings.

When the researchers looked at enterprise-grade, legitimate products designed explicitly to turn a novice developer into a competent coder, they found many aspiring cyber criminals experimenting with tools like ChatGPT and Claude, which the researchers said “excitedly report back on their discoveries”. However, Collier noted that a deeper exploration of these discussions found that, in most cases, forum members lacked the basic technical skills needed to use AI tools effectively for committing cyber crime.

“They’re using vibe coding tools for hobby projects, but particularly for the basic logistics of cyber crime operations,” he said. “Most of the coding involved in cyber crime isn’t hacking. It’s the same administration and basic engineering works that you’d need for any small startup, which means a lot of them don’t actually need to jailbreak Claude to get real utility out of it.”

The pessimistic view is that as these tools evolve, they will be able to be used for sophisticated cyber attacks. Adam Beaumont, interim director at the AI Security Institute (ASI), discussed the pessimist view. Beaumont, the former chief AI officer at GCHQ, said the ASI recently demonstrated how a frontier AI model executed a 32-step cyber attack against a simulated corporate environment from initial reconnaissance through to full network takeover.

“We estimate it would take a skilled human professional 20 hours’ worth of work, and this was the first time any model had done it, and weeks later, we tested a second model,” he said.

Beaumont pointed out that the attack he described was not a model answering a question about hacking. “It was a system that hacked,” he said. “We still don’t fully know how to ensure these systems act as we intend, or how to guarantee they remain under meaningful human control as they grow more capable.”

Beaumont called the ASI demonstration an “honest starting point”. “The uncertainty is real and the discomfort is appropriate,” he said.

For Beaumont, it represents something that can be built up to enable government, industry and the research community to make decisions based on what these systems can actually do built on evidence.