Back to news

January 7, 2019 | International, Aerospace, C4ISR

Pentagon Seeks a List of Ethical Principles for Using AI in War

BY PATRICK TUCKER

An advisory board is drafting guidelines that may help shape worldwide norms for military artificial intelligence — and woo Silicon Valley to defense work.

U.S. defense officials have asked the Defense Innovation Board for a set of ethical principles in the use of artificial intelligence in warfare. The principles are intended to guide a military whose interest in AI is accelerating — witness the new Joint Artificial Intelligence Center — and to reassure potential partners in Silicon Valley about how their AI products will be used.

Today, the primary document laying out what the military can and can't do with AI is a 2012 doctrine that says a human being must have veto power over any action an autonomous system might take in combat. It's brief, just four pages, and doesn't touch on any of the uses of AI for decision support, predictive analytics, etc. where players like Google, Microsoft, Amazon, and others are making fast strides in commercial environments.

“AI scientists have expressed concern about how DoD intends to use artificial intelligence. While the DoD has a policy on the role of autonomy in weapons, it currently lacks a broader policy on how it will use artificial intelligence across the broad range of military missions,” said Paul Scharre, the author of Army of None: Autonomous Weapons and the Future of War.

Josh Marcuse, executive director of the Defense Innovation Board, said crafting these principles will help the department “safely and responsibly” employ new technologies. “I think it's important when dealing with a field that's emergent to think through all the ramifications,” he said.

The Board, a group of Silicon Valley corporate and thought leaders chaired by former Google and Alphabet chairman Eric Schmidt, will make the list public at its June meeting. Defense Department leaders will take them under consideration.

Marcuse believes that the Pentagon can be a leader not just in employing AI but in establishing guidelines for safe use — just as the military pioneered safety standards for aviation. “The Department of Defense should lead in this area as we have with other technologies in the past. I want to make sure the department is not just leading in developing AI for military purposes but also in developing ethics to use AI in military purposes,” he says.

The effort, in part, is a response to what happened with the military's Project Maven, the Pentagon's flagship AI project with Google as its partner. The effort applied artificial intelligence to the vast store of video and and image footage that the Defense Department gathers to guide airstrikes. Defense officials emphasized repeatedly that the AI was intended only to cut down the workload of human analysts. But they also acknowledged that the ultimate goal was to help the military do what it does better, which sometimes means finding and killing humans. An employee revolt ensued at Google. Employees resigned en masse and the company said that they wouldn't renew the contract.

Scharre, who leads the Technology and National Security Program at the Center for a New American Security, said, “One of the challenges for things like Project Maven, which uses AItechnology to process drone video feeds, is that some scientists expressed concern about where the technology may be heading. A public set of AI principles will help clarify DoD's intentions regarding artificial intelligence.”

Full artcile: https://www.defenseone.com/technology/2019/01/pentagon-seeks-list-ethical-principles-using-ai-war/153940/

On the same subject

  • Raytheon Provides Cybersecurity for Global Hawk UAS

    January 24, 2019 | International, Aerospace, C4ISR

    Raytheon Provides Cybersecurity for Global Hawk UAS

    Mike Rees Raytheon Company has announced that it will deploy sustainment and cybersecurity experts around the world to support the ground control systems and onboard sensors used by the U.S. Air Force fleet of RQ-4 Global Hawk remotely piloted aircraft. Raytheon Intelligence, Information and Services will perform the work, which includes providing software upgrades to defend against cyber threats, as part of a $65 million subcontract from the aircraft manufacturer, Northrop Grumman. “Raytheon will help these unmanned aircraft meet tomorrow's threats,” said Todd Probert, vice president of Mission Support and Modernization at Raytheon IIS. “We have been improving the Global Hawk fleet's capabilities for 20 years by modernizing their ground and sensor systems and will now ensure their resiliency in the face of cyber threats.” Raytheon previously announced a $104 million effort to modernize the Global Hawk ground segment, moving payload and aircraft operators into mission control buildings. These new stations replace mission control, and launch and recovery elements previously housed in shelters. https://www.unmannedsystemstechnology.com/2019/01/raytheon-provides-cybersecurity-for-global-hawk-uas/

  • L'intelligence artificielle fait évoluer le paysage de la défense

    April 16, 2020 | International, C4ISR

    L'intelligence artificielle fait évoluer le paysage de la défense

    Une étude vient mettre en avant l'impact de l'intelligence artificielle dans les armées, à la fois sur un plan opérationnel et industriel. L'IA, facteur d'innovation. La révolution et les évolutions apportées par l'intelligence artificielle ne sont plus à démontrer. Dans le monde civil comme militaire, algorithmes et IA viennent apporter un complément d'information et représentent une plus-value, d'autant plus incontestable dans le domaine militaire. L'ensemble des armées du monde ont bien compris la nécessité de se tourner vers le big data et l'IA afin d'innover et de parvenir à conserver une supériorité opérationnelle sur l'adversaire. Les stratégies d'IA appliquées à la défense se multiplient et s'accompagnent généralement de réflexions sur les questions éthiques, dans un contexte occidental où l'emploi de SALA est vivement critiqué. Compétition internationale. Market Forecast a consacré une étude à ce sujet, et est venu analyser les impacts de l'IA au domaine de la Défense. C'est-à-dire aussi bien les apports pour les armées que la façon dont les travaux sont menés par les industriels. Nommé « Artificial Intelligence will arm defense contractors for global growth », ce rapport met en avant la façon dont les Etats se sont tournés vers cette nouvelle technologie afin de venir compléter les moyens de leurs armées. « Des nations comme la Chine ou la Russie réalisent des investissements substantiels dans l'IA pour des applications militaires, venant parfois poser des questions au regard du droit international et des droits humains », peut-on lire dans le rapport de Market Forecast. Des investissements visant ainsi à concurrencer les pays membres de l'OTAN, qui se sont eux aussi tournés largement vers l'intelligence artificielle afin de renforcer les capacités de leurs plateformes. Volet industriel. Mais l'adoption de l'IA dans les technologies de défense vient également questionner le tissu industriel. Selon Market Forecast, dans une grande majorité des cas, les Etats devront se tourner vers de nouvelles entreprises, plus petites et souvent non issues du secteur de la Défense afin de pouvoir se doter d'algorithmes performants. « Par exemple, aux Etats-Unis, plus de 60% des contrats tournés vers l'IA, inscrits dans le budget présidentiel de l'année fiscale 2017, sont menés par des acteurs non traditionnels de la Défense et des petits fournisseurs », rapporte Market Forecast. S'il n'y a aucun doute sur le fait que les géants de la défense et fournisseurs historiques continueront à mener leur activité, beaucoup s'appuieront donc sur de plus petites structures, spécialisées, afin d'intégrer la brique IA. En France, on peut ainsi penser à la petite entreprise EarthCube, qui malgré sa taille réduite et son jeune 'ge fait fortement parler d'elle et est parvenue à décrocher d'importants contrats dans la défense. Elle en a même fait son cœur d'activité. A l'avenir, les géants de la défense devront donc marcher aux côtés des plus petits et apprendre à collaborer avec ces structures qui portent elles aussi l'innovation. Car en effet, le marché de l'IA pour la défense devrait connaître un taux de croissance annuel moyen de 10,29% entre 2020 et 2028, démontrant l'importance de cette technologie qui offre d'importants débouchés commerciaux. https://www.air-cosmos.com/article/lintelligence-artificielle-fait-voluer-le-paysage-de-la-dfense-22924

  • Navy Wants Robot Boats But Will Still Need Sailors To Fix Them

    May 7, 2020 | International, Naval

    Navy Wants Robot Boats But Will Still Need Sailors To Fix Them

    "We need to find a balance of vehicle designs that enables the cost to be cheap enough that we can afford them, but it's not so highly optimized towards the purely unmanned spectrum that it's cost prohibitive to maintain them." By PAUL MCLEARYon May 06, 2020 at 3:27 PM WASHINGTON: The Navy needs to comb through a host of thorny issues before deploying a new fleet of unmanned ships to confront China, Russia, and Iran. “You don't hear me talking about artificial intelligence and machine learning and things like that just yet,” said Capt. Pete Small, the Navy's program manager for Unmanned Maritime Systems at a C4ISRnet conference this morning. “Those aren't my first concerns. My first concerns are about the field stability and sustainability of these systems right now.” The Navy just isn't equipped to deploy or sustain a new fleet of unmanned vessels yet. “Our infrastructure right now is optimized around manned warships,” Small said. “We're gonna have to shift that infrastructure for how we prepare, deploy, and transit” over large bodies of water before large numbers of unmanned vessels can be effective, he said. It's not clear where that planning stands, but the service has already invested tens of millions in the early work of developing a family of large and medium unmanned vessels, and is looking to vastly ramp that up in the 2021 budget, asking for $580 million for research and development. In 2019 an unmanned Sea Hunter prototype autonomous vessel sailed from San Diego to Hawaii, but it needed to repair several broken systems along the way, forcing sailors to board the ship. It was the first experiment of its kind, one the Navy has not repeated. Those mechanical problems point to work the Navy must do to reconfigure its logistics tail to meet the needs of a new class of ship. “We're going to have to transition from a [system] more optimized around our manned fleet infrastructure to a more distributed mix of these large manned platforms to smaller platforms,” Small said, “we're gonna need to talk about things like, tenders for heavy lift ships, or forward operating bases, things like that.” The early thinking is the service will use the ships as sensors deployed well forward of manned ships and carrier strike groups, which could be at risk if they maneuvered too close to contested waters. But the Navy isn't going to pin everything on a nascent fleet of robot boats — a new class of manned frigates is also being built to operate inside the range of enemy precision weapons. The frigates are going to be smaller and faster than current destroyers, with the ability to generate much more power so they can use lasers and other weapons for both offensive and defensive missions. The Navy is considering several sizes of USVs, including a large variant between 200 and 300 feet in length and having full load displacements of 1,000 tons to 2,000 tons. The ships should be low-cost, and reconfigurable with lots of room capacity for carrying various payloads, including mine hunting and anti-surface warfare. The 2021 budget submission proposes using research and development funding to acquire two more prototypes and another in 2022. Plans then call for buying deployable LUSVs at a rate of two per year. Medium unmanned ships will likely come in at between 45 to 190 feet long, with displacements of roughly 500 tons. The medium ships are thought to skew more toward mission modules revolving around intelligence, surveillance and reconnaissance payloads and electronic warfare systems. The first MUSV prototype was funded in 2019, and the Navy wants to fund a second prototype in 2023. Fundamental issues need to be sorted out before the Navy buys one of these ships. “We need to find a balance of vehicle designs that enables the cost to be cheap enough that we can afford them, but it's not so highly optimized towards the purely unmanned spectrum that it's cost prohibitive to maintain them,” Small said. If the maintenance is too complicated and time consuming, and “we have to take the whole vehicle out of the water and take it apart in some explicit manner to replace the parts, it's not gonna really support what we need in the field. So really, the sustainability of the technology is as important — if not more important — in the near-term than the technology itself.” https://breakingdefense.com/2020/05/navy-wants-robot-boats-but-will-still-need-sailors-to-fix-them/

All news