Artificial intelligence models developed using location scan data from the hugely popular augmented reality game Pokémon Go may soon play a role in assisting military drones operating in war zones. The technology is designed to enable drones to maintain precise navigation and coordination, even in environments where Global Positioning System (GPS) signals are unavailable, spoofed, or jammed.
Pokémon Go, launched in 2016 by Niantic in partnership with Nintendo, quickly became a global phenomenon, allowing players to find virtual creatures in real-world locations using their mobile phone cameras. A 2021 update introduced 'Pokéstops', offering in-game rewards to players who opted in to scan and upload recordings of physical locations. Niantic collected this user-generated spatial data, which was subsequently used to train its AI models to recognise and interpret real-world spaces.
Niantic recently confirmed a partnership with Vantor, a company specialising in spatial detection software for drones, including those utilised by various militaries. While both companies have stated that the raw ground scans from Pokémon Go were not directly provided to Vantor, the historical data was instrumental in training Niantic's foundational AI models. This agreement aims to provide a critical solution to a significant vulnerability in modern military operations, as stated in a December announcement from the companies.
The partnership's goal is to ensure autonomous systems and field teams can orient, coordinate, and maintain accurate situational awareness when satellite signals are compromised. Peter Wilczynski, Vantor's chief product officer, highlighted the benefits of combining their expertise in an interview, noting the need for rapid upgrades and integration of new hardware and software in the modern battle space.
However, the revelation has sparked concerns among privacy advocates. Tom Sulston, head of policy for the tech policy think tank Digital Rights Watch, described the use of civilian data for military ends as 'troubling'. He noted that while terms and conditions may include disclaimers, most users do not thoroughly read extensive legal documents when engaging with a video game. Sulston called for regulators to implement 'best interests of the user' or 'fair and reasonable' tests to safeguard users from such potential exploitation.
Dr Rob Nicholls, a senior researcher at the University of Sydney's centre for AI, trust and governance, suggested this case might be just the 'tip of the iceberg' concerning data collected from consumer applications being repurposed. He drew parallels to previous instances where data from fitness apps like Strava was used to identify military facilities, leading to directives from various militaries to avoid using GPS-enabled and sharing devices.