Jamayka Rhose Pascual

OpenAI has banned several ChatGPT accounts operating from the Philippines for generating coordinated social media content that supports President Ferdinand “Bongbong” Marcos Jr. and attacks Vice President Sara Duterte.

Photo Courtesy of Hugpong ng Pagbabago.

In its June 5 report, OpenAI identified Comm&Sense Inc., a marketing and public relations firm based in Makati, as the group behind this propaganda.

This marks one of the first publicly documented cases where a Philippine firm has been directly linked to an artificial intelligence (AI)-driven political influence campaign.

According to OpenAI, the banned accounts were part of a covert effort that used artificial intelligence to generate partisan messages designed to shape public opinion online.

The firm reportedly used ChatGPT to analyze political discourse, identify pro-Marcos narratives, and mass-produce short social media comments in both English and Filipino.

These comments, often fewer than 10 words, appeared crafted to blend into everyday social media discourses while subtly carrying a clear political agenda.

OpenAI found five TikTok channels tied to the operation began posting identical videos in mid-February, each with different captions to give the illusion of originality.

The comment sections are also filled with prewritten praises from inauthentic accounts with no followers and original posts.

On Facebook, the same strategy appeared in news posts from mainstream outlets such as ABS-CBN, using fake accounts.

These profiles lacked friends, timeline activities, signs of genuine engagement, and were found to have been created only around December 2024.

OpenAI said the campaign followed a methodical process of analyzing content, generating replies, and drafting promotional materials that exaggerated its effectiveness.

The promotional pitches claimed the campaign has achieved high reach and engagement, likely to impress current or potential clients.

Investigators named the operation “High Five” in reference to the upbeat tone, frequent emoji use, and pro-Marcos comments.

Although the campaign published a large volume of comments, most of them failed to attract likes, shares, and replies from real social media users.

OpenAI still flagged the campaign as a serious case of AI misuse aimed at distorting political discourses.

The company promptly banned the accounts involved and reaffirmed its commitment to preventing the use of AI tools for any propaganda.

As of writing, Comm&Sense Inc. has not yet responded to requests for a statement regarding the report.