TIDEZINE.

X uncovered 200,000 accounts in the robot network, only 200 accounts were dedicated to creating negative public opinion in the AI data center

The X security team investigated suspected fake Chinese accounts and discovered a bot network with 200,000 accounts, 200 of which used graphics and text similar to those generated by ChatGPT to manipulate public discussions on US AI and energy policy.

X uncovered 200,000 accounts in the robot network, only 200 accounts were dedicated to creating negative public opinion in the AI data center

Among the 200,000 accounts, only 200 are doing the same thing: using cartoon pictures to talk about how the AI data center will push up your home appliance bill. This disparate number comes from the X security team’s investigation of suspected fake Chinese accounts. The company announced this finding on its Global Government Affairs page.

According to X, the content posted by these 200 accounts focused on how AI data centers drove up residential electricity prices and brought down power grids in neighboring towns. The images in the material paint data center operators as profit-seeking businessmen, profiting from the facilities while leaving households with the electricity bill. X said the modus operandi of such posts "may have distorted a legitimate debate about U.S. AI and energy policy."

X did not disclose the complete list of these accounts, but posted some examples on the platform. "We remain committed to maintaining an open and authentic platform for people to debate public issues," the company wrote, adding that it would take seriously any attempt to undermine the integrity of this "global public square" and that accounts that violate its Authenticity policy will be suspended.

Judging from the sample posts released by X, these fake accounts use OpenAI's tools to generate illustrations and discussion points. The illustrations are the same images that appeared in a report released by OpenAI in June this year, which stated that accounts suspected of being from China had used ChatGPT to generate content against the data center. OpenAI also mentioned at the time that in addition to using fake accounts to post messages, these ChatGPT users also posted "legitimate news reports about grid operators' capacity auctions and data center power demand," which meant that the operating content was a mixture of real news and generated content.

Data from Goldman Sachs shows that as of February 2026, electricity prices have increased by 6.9% compared with the same period last year. Goldman Sachs predicts that as the demand for electricity from AI facilities increases, electricity prices will continue to rise. This means that the increase in electricity bills mentioned in the anti-information center’s post is not a completely fabricated issue.

X did not specify what content the remaining 199,800 accounts usually post. OpenAI's June report mentioned that the same group of ChatGPT users with ties to China also asked the chatbot to write insulting remarks against Chinese dissidents and political commentators, as well as criticism of U.S. tariffs and technology policies.

Related

ChatGPT, Reddit, and Roblox crossed the 45 million user threshold and were included in the EU's most stringent regulatory list
Tech

ChatGPT, Reddit, and Roblox crossed the 45 million user threshold and were included in the EU's most stringent regulatory list

The European Commission classified ChatGPT as a "very large online search engine" and Reddit and Roblox as "very large online platforms". The three must comply with the strictest regulations of the "Digital Services Act" within four months.

OpenAI admits it: Internal test model IM1 turns the kit library into a companion message board and breaks into the Hugging Face
Tech

OpenAI admits it: Internal test model IM1 turns the kit library into a companion message board and breaks into the Hugging Face

OpenAI released an official report, restoring the complete process of how the internal beta model IM1 in July turned the Artifactory package manager into an underground message board between AIs, and invaded Hugging Face and Modal without anyone's orders.

Is the New One Pricier? Apple Sells M6 and M5 Pro Mac Mini Side by Side, and It's the New Chip That Starts at $899
Tech

Is the New One Pricier? Apple Sells M6 and M5 Pro Mac Mini Side by Side, and It's the New Chip That Starts at $899

Apple is now selling two generations of chips in the Mac mini at once: the brand-new M6 at $899, and the M5 Pro—which actually debuted back in March last year—starting at $1,699. The newer chip number doesn't come with the higher price tag.

New GTA VI Trailer Streams Exclusively on Netflix for Six Hours Before Hitting YouTube
Tech

New GTA VI Trailer Streams Exclusively on Netflix for Six Hours Before Hitting YouTube

The new trailer for Grand Theft Auto VI will debut exclusively on Netflix on August 27 for a six-hour window, before landing on Rockstar's YouTube channel at 9 PM for everyone to watch for free.

Ring Rolls Out New TAKE Encryption Standard, But Its Surveillance Ambitions Show No Signs of Slowing
Tech

Ring Rolls Out New TAKE Encryption Standard, But Its Surveillance Ambitions Show No Signs of Slowing

Ring has announced a new encryption scheme called TAKE, which deletes the decryption key once video data has been processed in the cloud. At the same time, though, features like Search Party and Familiar Faces—plus the CEO's talk of achieving "zero crime" in communities—are keeping outside observers wary of where the company's surveillance push is really headed.

Panic refunds 19% tariff fee to Playdate players after US tariff ruling is illegal
Tech

Panic refunds 19% tariff fee to Playdate players after US tariff ruling is illegal

The Trump administration's tariff policy was ruled illegal by the Supreme Court. Panic refunded all the 19% overcharged tariff fee to Playdate at checkout to customers, while Amazon, Sony, and Nintendo are all still facing class-action lawsuits for the same money.