OpenAI Bans Philippine Accounts Over AI-Generated Political Content
2 min read
JUNE 10,2025
MANILA, Philippines — OpenAI, the U.S.-based company behind ChatGPT, has banned several Philippine accounts for using its AI to create political content designed to sway public opinion on social media. The banned accounts supported President Ferdinand Marcos Jr. and criticized Vice President Sara Duterte. This was revealed in OpenAI’s June 5, 2025 report titled “Disrupting Malicious Uses of AI.”
Why Did OpenAI Ban the Accounts?
OpenAI found that certain ChatGPT accounts generated a large volume of short comments in English and Taglish. These comments, posted on TikTok and Facebook, focused on Philippine politics and current events.
“We banned ChatGPT accounts that used our models to generate bulk volumes of short comments,” the company stated. “These were posted on TikTok and Facebook and targeted political topics in the Philippines.”
OpenAI Bans Who Was Behind the Operation?
OpenAI identified Comm&Sense Inc., a Makati-based PR and marketing firm, as a key player in the campaign. As of this writing, the company has not responded to media inquiries from the Philippine Daily Inquirer.
What Was ‘Operation High Five’?
OpenAI labeled the influence campaign as “Operation High Five” due to its use of emojis and overly positive messaging. The comments consistently praised President Marcos and criticized Vice President Duterte.
Despite the effort, OpenAI classified it as a low-impact Category 2 operation on its six-level Influence Operation (IO) impact scale.
How Did the AI Campaign Work?
OpenAI’s investigation revealed the operation used ChatGPT in multiple stages:
- Trend Analysis: Operators used ChatGPT to scan and analyze posts related to Philippine politics, especially those about President Marcos.
- Comment Generation: ChatGPT created short, pro-Marcos comments — usually under 10 words.
- PR Creation: The tool helped draft public relations pitches and data analyses to attract clients. These pitches highlighted the campaign’s five TikTok channels promoting pro-administration content, launched in mid-February during the 2025 midterm election campaign.
How the Content Spread Online
The campaign spread AI-generated comments in two major ways:
- TikTok: Dozens of new accounts commented on videos to boost perceived popularity.
- Facebook: Similar short comments were posted under news articles from outlets like ABS-CBN News. Most accounts had minimal history and were likely created around December 2024.
Was the Campaign Effective?
According to OpenAI, the campaign had limited success:
- An ABS-CBN News post received hundreds of AI-generated comments but was buried under 23,000 organic replies.
- Some TikTok videos hit over 1 million views, while others barely reached 100.
- Most AI-generated comments received little to no engagement — few likes, shares, or replies.
OpenAI Bans Platform Responses
Facebook and TikTok both ban fake engagement and artificial boosting. Although this operation had limited reach, it highlights growing attempts to manipulate voters using AI during sensitive political seasons.
Why It Matters
With the 2025 Philippine midterm elections nearing, AI’s role in political messaging is under scrutiny. The incident underscores the need for stricter platform oversight and transparency as AI becomes more embedded in digital communication. NOWTREND
About The Author
