Investigation reveals AI is being used to smear safety advocates

In February, PauseAI received an email from journalist Michael Chen. Only Michael Chen does not exist.

In February, PauseAI received an email from Michael Chen, a journalist at The Wire by Acutus. He was looking for a response to an article he was writing about what he called “escalating anti-AI radicalism.”

Only Michael Chen does not exist.

According to an in-depth investigation by Model Republic, Michael Chen is an AI agent – an AI system capable of researching, emailing, interviewing, writing and publishing content.

Model Republic’s reporting indicates that almost the entire Acutus website is generated by AI and that the trail of ownership appears to lead back to OpenAI.

AI posing as a human

Acutus describes its content as “expert-sourced journalism” offering “independent reporting” from “expert voices from across industries.” Model Republic found that of the publication’s 94 articles, only three appear to have been written by humans. Acutus’ own internal records, which are publicly accessible, refer to an “AI interviewer”, a “reporter agent” and an “AI editorial review” process. Many quotes in Acutus articles appear to be copied from previously published web content, a hallmark of AI-produced text.

In other words: Michael Chen is a fake journalist, working for a fake publication.

Photo by Sivani Bandaru on Unsplash

A smear campaign that appears to lead to OpenAI

Model Republic says that more than a third of Acutus’ articles read like paid advocacy. This includes favourable coverage of the pharmaceutical industry, the cryptocurrency lobby, the natural-gas and data-centre lobbies and several 2026 Republican Senate campaigns.

Many of the site’s articles cover the topic of AI and echo anti-AI-regulation sentiments. The site criticises US states that regulate AI and was quick to condemn PauseAI following the recent attack on Sam Altman’s home, framing AI safety advocates as engaging in “anti-technology extremism.”

Model Republic’s exposé points to OpenAI as the likely backer of Acutus. This would not be the first time the company has been linked to a covert influencing campaign.

This tactic is known as astroturfing: hiding the sponsors of a message to make it appear as though it originates from ordinary citizens or grassroots organisations.

AI used to push a political agenda

Companies have always used PR and lobbying to shape public opinion. What is new is the scale and speed at which AI now allows them to do it. AI makes it easier than ever for these kinds of campaigns to go undetected, influencing individuals on a large scale, and ultimately shaping politics, culture and society.

The AI companies developing these powerful systems are the ones with the most to lose from regulation. And now it seems they are using the AI models they have created to smear the AI safety advocates asking them to slow down.

Acutus signals the insidious influence of AI companies and the level to which they are willing to stoop to silence their critics and beat their competitors in the race to build superintelligent systems. Perhaps the most worrying omen, though, is that a new generation of even more powerful, even more capable AI models is just around the corner. With the capabilities of AI systems doubling every six months or so, the next model will be even more adept at influencing public discourse undetected.

If OpenAI is behind Acutus, it does suggest – as many AI safety advocates continue to insist – that the race to develop ever-more powerful AI systems is accelerating out of control.

Public support for an AI pause is growing

The development of AI – systems that will likely soon be more capable than humans across the board – is in the hands of a small number of CEOs. This technology is already having an impact on the economy, the jobs market, education and mental health. This impact multiplies with each new model. No one voted for this.

On the contrary, a growing number of concerned citizens is demanding a pause in the development of the most powerful AI systems until we know how to build them safely and keep them under democratic control.

In the US, 64 percent feel that superhuman AI should not be developed until it is proven safe and controllable, or should never be developed, while there is overwhelming support (73 percent) for robust AI regulation.

As early as 2023, long before the emergence of the powerful AI systems we see today, a YouGov poll found 74 percent of the UK public agree that “preventing AI from quickly reaching superhuman capabilities” should be a priority, while 60 percent support a global treaty to ban the development of AI that is more intelligent than humans.

Even Anthropic, one of the largest AI companies in the world, has said:

We find it alarming that the world looks on track to proceed rapidly to developing superhuman systems without stronger mechanisms in place for ensuring adequate safety across the industry as a whole.

Professor Stuart Russell, author of the AI textbook used in universities around the world, has warned that it may take a Chernobyl-scale AI disaster for policymakers to act.

How many warnings do we need? Let’s not wait for a disaster.

Find out about our latest campaign and let us know about your interactions with AI here.

Watch Professor Stuart Russell tell MEPs in Brussels, “We are on a trajectory towards a loss of control.”*

Find out more about PauseAI