On 14 October 2023, the Philippines’ Division of Nationwide Protection (DND) issued a memorandum banning using synthetic intelligence (AI) picture generator apps by uniformed army personnel and DND civil servants. On the time, using AI picture apps similar to Remini and Photograph Lab went viral, with many Filipinos utilizing them to create different profile pictures for social media.
Whereas this looks like an overreaction to a innocent social media pattern, the safety implications had been critical. The rationale for the ban was to cease personnel from supplying picture knowledge that could possibly be used to ‘create faux profiles that may result in identification theft, social engineering, phishing assaults and different malicious actions’, together with deep fakes or AI-generated media utilizing pictures of actual individuals. This expertise’s dangerous potential has lengthy been recognised as it may be used to undermine belief in establishments and diplomacy and jeopardise nationwide safety by enhancing army deception, bolstering propaganda efforts and impeding intelligence operations.
There’s a threat that deep faux proliferation will trigger audiences — conscious that AI content material is spreading — to mistrust every little thing they see. This problem is problematic when it impacts the general public, however utilizing AI-generated impersonations to sow mistrust in army chains of command could be detrimental for morale, credibility and army effectiveness. Whereas primarily seen as a knowledge privateness and safety risk, AI software program has the potential to disrupt ‘upstream’ army duties — duties that happen prior to focus on choice and engagement — forcing slowdowns in army assist operations or affecting decision-making at important moments.
For the Philippines, this risk was demonstrated in Could 2024. Chinese language diplomats alleged that the then-commander of the Philippines’ Western Command had agreed to a ‘new mannequin’ for managing interactions in the course of the resupply of the BRP Sierra Madre within the Ayungin Shoal, often known as Second Thomas Shoal, releasing transcripts of the dialog. The response from the DND and the Armed Forces of the Philippines was swift, denouncing the transcript as a part of a ‘deep faux’ operation and manipulation of what the previous commander of the Western Command claimed was an off-the-cuff dialog with no point out of a brand new mannequin.
It appears that evidently the Chinese language efforts meant to trigger confusion and coerce the Philippine authorities into complying with a questionable settlement, claimed to have been made with the Duterte administration. Whatever the reality, this demonstrated the perils of adversaries gaining access to army officers’ audio.
This isn’t the primary time common apps have intersected with army operations — operating utility Strava revealed areas of United States and different army bases in 2018, whereas relationship apps similar to Tinder had been used to collect operationally related info in the course of the Russian invasion of Ukraine in 2022 by Ukrainian girls and sympathisers, with the identical occurring on the Russian aspect.
Whereas these examples noticed apps getting used as knowledge assortment instruments, extensively out there AI content material turbines characterize an evolution of expertise, combining fabrication and assortment capabilities.
The twin-use nature, accessibility and the variety of apps characterize huge troves of knowledge and massive challenges for regulation. With virtually 9 million smartphone apps out there on app shops and 225 billion downloads as of April 2024, guaranteeing that these apps and their mum or dad corporations safe the info they acquire and aren’t misused by firms or governments might be a problem.
As a substitute of wider regulation, directives such because the October 2023 DND memo are helpful for producing consciousness of the dangers posed by AI content material turbines. Good ‘cyber hygiene’ practices are essential in trendy militaries and are important in minimising the opportunity of knowledge being compromised. The fact of AI picture turbines may additionally require army officers to think about enhancing safety of their likenesses, just like persona and publicity rights, or instituting protocols to minimise threat of their likeness being acquired for nefarious functions.
In the long term, militaries might want to improve media literacy, disinformation training and AI content material recognition expertise. Strengthening army skilled and civic training might be essential. Recognising deep faux content material may additionally be inadequate if the focused personnel are predisposed to believing the messages conveyed by deep fakes.
Defending army info areas from being undermined by adversaries will stay a permanent problem for militaries internationally, together with the Philippines. Militaries might want to steadiness important pondering, technological savvy and institutional procedures to make sure that they’ll proceed to function successfully in an more and more fraught info surroundings.
Erick Nielson C Javier is a Defence Analysis Officer on the Nationwide Protection School of the Philippines.
