Carefully, deliberately, cautiously, US DHS is embracing artificial intelligence (AI) and machine learning (ML) technology. But a host of relatively new early-stage companies are not waiting for DHS to develop a final AI “plan” – they’re developing a range of AI-powered technologies with homeland security applications, some of which are already being used by DHS component agencies such as CBP and TSA.
DHS’ AI Strategic Planning and Implementation Process
In December 2020, DHS published its AI and ML Strategy.
The document is notable for its emphasis on the risks posed by AI. The first sentence of DHS’ “strategic vision” for AI reflects this perspective:
“Our strategic vision is that DHS will become a global leader in policy development, governance, and the use of AI systems as we lead national efforts to mitigate against institutional and enterprise risks posed by AI.” [my italics added][1]
DHS has said AI presents opportunities “to more effectively or efficiently accomplish our mission to secure the homeland. Yet with increased use of AI systems across the homeland security enterprise, comes increased risk. These risks include compromised or poorly designed AI systems, as well as adversarial use of AI technologies by unfriendly nations or criminals to increase their malicious capabilities.”
And a significant portion of the document and DHS’s strategy is focused on mitigating the potential risks posed by AI: the risk that AI systems will be poorly designed, and produce bad results, the risk that AI systems could be compromised by bad actors or adversaries, and the risk that there will be a public backlash against any use of AI by DHS. The approach DHS is taking to implement AI is to ensure that AI is being used in a way that is “respectful of the values of the American people and assess the impact of Artificial Intelligence on the Homeland Security Enterprise, invest in AI, assess the risks to the country and department, and lastly develop a task force based around AI while improving public trust and engagement.”
The DHS strategy identifies only one specific “positive” application of AI for homeland security – the use of AI to help improve cybersecurity. Other potential applications of AI are left to be determined and identified through a soon to be released “DHS AI Implementation Strategy”.
NeuralGuard and Shield AI
Meanwhile, innovative technology companies are already developing applications for the use of AI to improve homeland security and national defense.
Neural Guard’s mission is to utilize artificial intelligence to perfect the threat detection process during the x-ray security screening process. NeuralGuard software automatically detects potential threats in the x-ray image that can be difficult for an x-ray operator to spot on their own. NeuralGuard is not yet being used by DHS, but the company has sold its solution to other US government agencies and to customers worldwide.
And while DHS develops its AI strategy, companies such as Shield AI, whose mission is to protect service members and civilians with intelligent systems, are using AI technology to build fully autonomous aviation to help the US Defense Department.
According to a recent company press release, the Shield AI’s Hivemind software is an AI pilot for military and commercial aircraft that enables intelligent teams of aircraft to perform missions ranging from room clearance, to penetrating air defense systems, and dogfighting F-16s. Hivemind employs state-of-the-art algorithms for planning, mapping, and state-estimation to enable aircraft to execute dynamic flight maneuvers and uses reinforcement learning for discovery, learning, and execution of winning tactics and strategies. On aircraft, Hivemind enables full autonomy and is designed to run fully on the edge, disconnected from the cloud, in high threat, GPS and communication-degraded environments.
Conclusion
NeuralGuard and Shield AI demonstrate that there are exciting and powerful AI applications available to support national security – let’s hope DHS leadership finalizes it’s AI planning so that it can take advantage of them.
[1] https://www.dhs.gov/sites/default/files/publications/dhs_ai_strategy.pdf
