In early October, following an OpenAI investigation, we learned that CounterPunch had again been catfished by a fake persona, this time allegedly linked to an Iran-backed operation.

The story goes that Iran used OpenAI agents to write and submit articles to U.S. media outlets. Many were affected by what OpenAI has dubbed “Bogus Bylines.” Other outlets that published these AI deepfakes include Common Dreams, Middle East Monitor, LA Progressive, DailyKos, and Modern Diplomacy, among others.

In our case, two writers, Sophia Gonzalez and Ervin Hoskins, were published several times this past summer. Their articles, largely unexceptional, covered a range of topics: ICE, Trump, and U.S. policy in the Middle East.

None of the articles were directly submitted to us by these fake authors. Instead, they landed in our inbox via PeaceVoice, a well-respected syndication service run by Tom Hastings.

Subsequently, we’ve deleted every article by both of these AI-generated writers.

Following the infamous Alice Donovan affair (a fake writer allegedly tied to Russia), CounterPunch instituted a stringent protocol for new authors who weren’t easily verifiable. Yet we were lax in verifying writers submitted to us through syndication services such as PeaceVoice.

Tom Hastings has not issued a formal statement about the “Bogus Bylines” operation that impacted his service and many outlets, but he has confirmed that he is instituting a stronger AI policy.

Here at CounterPunch, we maintain a strict no-AI rule for content creation, though monitoring writers’ use of AI tools is admittedly difficult. Even so, we’re fully committed to ensuring that every word you read on CounterPunch is written by a human.

We apologize for the oversight.