AIBRIEF
US defense AI firm Shield AI and Taiwan's defense research institute (NCSIST) ran NCSIST's indigenous Mighty Hornet III drones through one Taipei search, with Hivemind autonomy software dividing the area and landing them all.
Shield AI, a US defense-autonomy firm, and Taiwan's National Chung-Shan Institute of Science and Technology (NCSIST) flew three Mighty Hornet III drones together in Taipei on August 6, with Hivemind, one autonomy software, deciding where each drone searched (Shield AI press release).
That is the practical meaning of "cooperative autonomous search" in this release. Hivemind launched the three aircraft, divided the search area into separate sub-areas, kept one drone clear of a partial no-fly zone inside its sector, and landed all three without a human pilot steering any of them. The software was flying; no human was.
This was a demonstration, not an operational deployment. NCSIST engineers built the indigenous AI pilots in under three months, with Shield AI support, on the institute's own Mighty Hornet III airframe (Taiwan News).
Shield AI President Brandon Tseng said autonomy is what lets any country fly large numbers of drones, and that helping Taiwan develop AI pilots would "significantly enhance the nation's defensive capabilities." NCSIST President Li Shih-chiang said the institute can now operate Hivemind on its own.
The Taipei flight followed Shield AI's July maritime demo in Taiwan with Thunder Tiger, where Hivemind coordinated uncrewed surface vessels (Naval News). Follow-on coverage notes the renewed and expanded deal (Defence Blog).
What the public record does not yet show is whether the three-month integration includes adaptive behavior beyond the announced sub-area assignments and geofence respect.