MANILA, Philippines — All defense and military personnel in the Philippines have been ordered to refrain from using all artificial intelligence image generator apps, which a government memorandum said poses “significant privacy and security risks.”
In an October 14 issued by Defense Secretary Gilbert Teodoro, the defense chief warned of the “inherent security risks” posed by uploading real images of one’s face to reproduce AI images — technology that has also triggered global concerns on how generative AI can be misused to create deceptive or abusive content.
"The online trending digital application that uses AI which requires its users to submit at least 10 photos to generate an enhanced portrait, poses significant privacy and security risks," read the memorandum, which covers all personnel from the Armed Forces of the Philippines and the Department of National Defense.
The memorandum added that the AI image generator compiles users' data and creates a "digital person that mimics how a real individual speaks and moves."
Teodoro said that while "seemingly harmless and amusing," AI-powered applications can be used to create fake profiles that can lead to "identity theft, social engineering, phishing attacks, and other malicious activities.”
Without elaborating, Teodoro said that there has "already been a report of such a case."
The defense chief also called on all DND and AFP personnel to “practice vigilance in sharing information online.”
“Ensure that your actions are aligned to the Department's values and are in adherence to existing policies," Teodoro added.
Dangers of AI image generators
Artificial intelligence-focused magazine Analytics Insight warned in March that AI image generators can be misused to create false content that will fuel the rapid spread of online disinformation.
"Malicious actors" can use the data uploaded into AI image generators to create fake but realistic-looking social media profiles to scam people. Scammers can also use AI to create realistic images and solicit donations when devastating natural disasters strike, Analytics Insight said.
Analytics Insight also highlighted how AI-generated photos cannot be found via reverse image search and are more difficult to verify.
"The cybercriminal does not have to work with a limited number of photographs to deceive their target-by using AI, they can generate as many as they want, creating a convincing online identity from scratch," the AI-focused magazine said.
The Philippines has yet to pass legislation that will lay down rules to protect the public from risks posed by certain types of AI technologies, but a House bill (pushing for such protections (House Bill 7396) is currently pending at the committee level.