OpenAI Bans Pro-Marcos Jr AI Accounts, Exposing Philippines' Misinformation Vulnerability
MANILA, Philippines – OpenAI, has banned accounts in the Philippines that used ChatGPT to generate social media comments favoring President Ferdinand Marcos Jr. and criticizing Vice President Sara Duterte. This operation, named "Operation High Five," involved creating partisan content for platforms like TikTok and Facebook and has been linked to Comm&Sense Inc., a Makati-based public relations and marketing company. This incident, detailed in OpenAI's report "Disrupting malicious uses of AI," highlights the growing threat of AI-driven political influence operations, especially with the Philippines' 2025 midterm elections.
This development indicates a shift where AI is used for precise, domestic political manipulation, influencing narratives and potentially electoral outcomes. The alleged involvement of a recognized PR firm in a covert influence operation suggests legitimate communication tools are being co-opted for deceptive purposes, making it harder for the public to distinguish authentic from manufactured online discourse. OpenAI's decisive action sets a precedent for corporate responsibility, emphasizing tech companies' role in defending information integrity.
"Operation High Five" was a sophisticated, multi-platform influence campaign aimed at artificially inflating support for President Marcos Jr. and undermining Vice President Sara Duterte. Its primary goals were to generate social media comments praising Marcos and his initiatives, while criticizing Duterte, even nicknaming her "Princess Fiona."
The broader aim was to "inundate" comment sections and create a false impression of popularity for associated TikTok channels. Tactics included using ChatGPT to analyze social media posts for themes, then generating "bulk volumes of short comments" (under 10 words) in English and Taglish. The campaign set up five TikTok channels with identical videos, which dozens of anonymous TikTok accounts then flooded with comments to simulate widespread support.
Similar AI-generated comments were posted on Facebook under news reports from legitimate outlets like ABS-CBN, often by newly created accounts with minimal friends.
ChatGPT was also used to generate "public relations pitches and statistical analyses" for the operation itself, possibly for clients. The TikTok channels began posting in mid-February, coinciding with the 2025 midterm election campaign period. OpenAI's investigation explicitly tagged Comm&Sense Inc., a Makati-based PR and marketing company owned by Jaeger Tanco, as being among the entities behind this campaign.
Despite the volume of content, OpenAI assessed "Operation High Five" as a "low-impact Category 2 activity" (on a scale of 1-6), noting "little engagement from real users." Most comments received single-digit or zero replies, likes, or shares. OpenAI's rapid detection and banning of these accounts likely mitigated their potential reach.
The "low-impact" classification doesn't diminish the intent or sophistication. The tactics, like creating fake accounts to "simulate popularity," reveal a strategic understanding by threat actors to simulate authentic online interaction and social proof, exploiting human biases. The use of ChatGPT for "public relations pitches" suggests a commercial "disinformation-as-a-service" model, making sophisticated AI campaigns accessible to more actors.
"Operation High Five" is part of OpenAI's broader global campaign against the malicious use of its AI tools. OpenAI's June 5 report, "Disrupting malicious uses of AI," detailed 10 distinct operations disrupted recently, spanning multiple countries including China, Russia, Iran, North Korea, Cambodia, and the Philippines. Malicious uses of ChatGPT varied, including social media comment generation, malware refinement, cyberattack assistance, and employment scams.
OpenAI's usage policies explicitly prohibit generating disinformation, misinformation, false online engagement, and engaging in political campaigning or lobbying. To counter these threats, OpenAI uses automated systems, human review, and user reports. The company leverages AI as a "force multiplier" for its investigative teams to detect, disrupt, and expose abusive activities. Each disruption provides insights to refine defenses, and OpenAI shares indicators with industry partners.
OpenAI acknowledges a continuous "arms race" where malicious actors innovate AI exploitation, while AI companies use AI to counter these threats. This dynamic requires sustained investment in AI safety and security. OpenAI's detailed policies and proactive disruption reports signal a growing ethical commitment to safeguarding information integrity and democratic processes.
