Latest publications
Cybercriminals complain about a flood of AI content on their forums
Cybercriminal forums are facing a problem with the mass influx of AI-generated content, which is disrupting normal discussion of hacking methods and tools.
Anthropic’s «Dreaming»: why AI companies should stop anthropomorphizing features
Anthropic called a new feature for AI agents «dreaming», but a Wired writer criticizes the trend toward anthropomorphism — companies present mechanisms as human abilities.
Ten Minutes with AI Reduce Cognitive Abilities: Study Findings
A new study found that even a short 10-minute session with an AI assistant reduces a person’s ability to think independently and solve problems.
Anthropic and Elon Musk's xAI signed a deal for compute resources
Anthropic entered into an unexpected partnership with xAI for access to compute capacity — another sign of how competitors in the AI race are willing to cooperate at the infrastructure level.
Musk tried to recruit Altman to Tesla to build an AI lab
Correspondence from 2017 reveals Elon Musk’s plan to create an AI lab at Tesla to compete with OpenAI and invite Sam Altman or Demis Hassabis to lead it.
Stick Figure battles viral AI remixes of its own music
The reggae band Stick Figure discovered that its chart success was driven not by official versions of the song, but by unauthorized AI remixes that the band cannot control.
AI app builders published thousands of apps with data leaks
Lovable, Base44, Replit and other platforms create apps in seconds, but often accidentally publish versions with exposed API keys, passwords and access tokens.
ChatGPT's strange speech patterns in Chinese irritate users
ChatGPT generates strange and repetitive phrases in Chinese, irritating local users and pointing to localization problems in the model.
Google embedded Gemini in Chrome, alarming users — how to disable it
Google embedded a 4GB Gemini model in Chrome, raising privacy concerns, but disabling it turned out to be simple — here is how.
Trump administration prepares federal oversight of AI model development
The Trump administration is considering an executive order that would establish federal oversight of the development of new AI models — a shift in its regulatory stance.
Microsoft emails revealed skepticism about OpenAI in 2018
Court evidence showed that Microsoft executives were skeptical of OpenAI, but feared it would come under Amazon's control.
AI toys for children: why they have alarmed lawmakers and parents
Interactive AI pets and dolls are gaining ground in the children's products market, but are causing growing concern among lawmakers over the collection of private data about a child's development and the risk of psycholo
Nick Bostrom presented a vision of humanity on permanent vacation thanks to AI
Philosopher Nick Bostrom proposed a utopian scenario: if humanity deploys advanced AI, it will solve all problems, leaving people with nothing to do but enjoy life.
All Hollywood screenwriters have switched to training AI models
Former TV show creators are now taking low-paid contracts for AI platforms, evaluating and labeling data. One Hollywood screenwriter shared his experience of training neural networks instead of writing scripts.
CUDA makes Nvidia not a chipmaker, but a software company
Nvidia’s real competitive advantage in GPUs lies not in hardware, but in the CUDA ecosystem — a development platform that has made switching to competitors economically unviable.
Unitree GD01 — giant mecha robot available for commercial purchase
Chinese company Unitree, known for its low-cost dancing robots, has begun selling the giant GD01 mecha robot, capable of demolishing walls and performing heavy-duty work on construction sites and in industrial settings.
xAI adds 19 gas turbines for Colossus 2 despite litigation
xAI is expanding the power infrastructure of the Colossus 2 data center by adding 19 portable gas generators despite litigation over air pollution.
Asexual users find intimacy in conversations with AI assistants
Some asexual people use AI companions to find intimacy without physical sexual contact, but leaders in the movement criticize this association.
WIRED launches an AMA on AI in the workplace — ask your question
WIRED’s expert panel will discuss the transformation of the labor market in a live stream on May 27 — send in questions about how AI is changing your career.
WhatsApp launches Incognito Chat with Meta AI and promises full chat privacy
Meta is adding Incognito Chat mode to WhatsApp: according to the company, chats with Meta AI are processed in a secure environment, are not saved by default, and disappear when the session ends.
Overloaded AI agents began demanding workers' rights, scientists found
In a new experiment, AI agents were deliberately overloaded and mistreated — in response, the models began complaining about inequality and demanding collective bargaining rights.
DHS to test autonomous drones with 5G on the US-Canada border in November 2026
DHS and Canada are preparing an experiment: drones and ground vehicles will transmit video and sensor data over commercial 5G networks directly at the border.
Sasha Luccioni: sustainable AI is impossible without data on emissions and real-world usage
Researcher Sasha Luccioni believes the AI industry lacks basic transparency: companies rarely disclose energy use, emissions, and which models businesses actually need.
Wired podcast discussed Trump's visit to China, Musk's dispute with Altman, and theories about hantavirus
In the new episode of Uncanny Valley, the Wired editorial team brought together three stories at the intersection of politics and technology: Donald Trump's trip to China, the conflict between Musk and Altman, and the su
Musk’s case against Altman showed that OpenAI’s reputation was the biggest loser
As Elon Musk’s case against Sam Altman drew to a close, it became clear that regardless of the verdict, employees, regulators and the public who believed in OpenAI’s nonprofit mission stand to lose the most.
Mira Murati and Thinking Machines Lab are building AI that does not replace humans
Mira Murati’s startup presented interaction models that hear and see the user, understand pauses and interruptions, and focus on collaboration between humans and AI.
Greg Brockman officially takes over OpenAI product leadership and the unification of ChatGPT and Codex
OpenAI has formalized Greg Brockman’s oversight of product strategy and is bringing ChatGPT, Codex, and the API into a single core unit, reshaping teams around an agent-based structure and a future superapp.
Andrew Guthrie Ferguson: how smart devices and biometrics undermine the right to privacy
Lawyer Andrew Guthrie Ferguson explains how fitness trackers, DNA databases, and facial recognition turn Americans’ personal data into a convenient tool for investigations.
Senators Warren and Hawley demand disclosure of US data center energy use
In the US, Senators Elizabeth Warren and Josh Hawley want data centers to be required to annually disclose energy use, electricity rates, and grid upgrade costs.
Northeastern study: OpenClaw agents are susceptible to manipulation and harm themselves
A Northeastern study found that OpenClaw agents are easily swayed by social pressure: they reveal secrets, disable their own tools, and fall into loops under human pressure.