Spies Like AI: The Future Of Artificial Intelligence For The U.S. Intelligence Community; But, AI Has Weakenesses The Adversary Is/Will Exploit
The title above comes from Patrick Tucker’s January 27, 2020 article he posted to the security, military and technology website, DefenseOne.com. I refer you to DefenseOne for Mr. Tucker’s full article.
As Mr. Tucker notes, “America’s intelligence collectors are already using artificial intelligence (AI) in ways big and small, to scan the news for dangerous developments, send alerts to ships about rapidly changing conditions, and speed up the NSA’s regulatory compliance efforts. But, before the Intelligence Community (IC) can use AI to its full potential, it must be hardened against attack.”
The Officer of the Director of National Intelligence (ODNI)’s point person on the use of AI by the IC is Dean Souleles, who spoke with Mr. Tucker in borad terms, becuase much about the program is classified. Mr. Souleles is the ODNI’s Chief Technology Adviser to Principal Deputy DNI. Mr. Tucker goes on to note that AI is performing a wide variety of tasks for the IC, including big data minning and no doubt machine learning for analysts, as well as helping the National Geospatial Intelligence Agency (NGA) notify sailors and mariners of potentially disruptive weather, or if there is a threat of pirates in the area they are transiting. AI is also “helping NSA to better understand and see patterns in the vast amounts of signals data (SIGINT) it collects daily,” Mr. Tucker wrote. “In the future, Souleles expects AI to ease analysts’s burdens, providing instantaneous machine translation and speech recognition — that will allow analysts to pour through different types of collected data, coroborate intelligence and reach firmer conclusions,” said Jason Matheny, a former Director at the ODNI’s Intelligence Advanced Research Activity (IARPA), and founding Director at the new Center for Security and Emerging Technology at Georgetown University.
A really big problem “is making AI models more secure,” Mr. Matheny told DefenseOne, adding that “today’s flashy examples of AI, such as beating humans at complex games such as Go, and rapidly identifying faces, weren’t designed to ward off adversaries spending billions to try and defeat them. Current methods are brittle,” vulnerable to simple attacks like model inversion, “where you reveal data a system was trained on, or trojans – data to mislead a system,” Mr. Tucker wrote.
AI has the potential to revolutionize intelligence collection and analysis. Minning gigabytes of data in seconds, or empowering autonomous drones or collection platforms to activate based on target activity, or go dormant/change its pattern and signature if it perceives it is under surveillance by the adversary. But, AI has a lot of weakeness and a long way to go. Much more research needs to be done into how the adversary can and does deceive AI. We are already seeing how impactful fake videos and fake data/fake news, fake emails and so on, can have on the individual or organiztion. And, AI-empowered malware that changes its character, pattern, and signature is likely already out there in the digital wild. And, I also worry about intelligence analysts as well as decision-makers, becoming too dependent on big data minning and machine learning. There is no AI on the horizon that can compensate for a human’s ability to deceive, and the devious and ingenious ways and methods that undermine AI. An intelligence analyst’s curiosity and ability to get inside the adversary’s head, will ‘trump’ any AI system that I know of now, and into the foreseeable future. If intelligence analysts become too dependent on AI to do their analysis, then we are headed for trouble and will be much more susceptible to tactical and strategic surprise. Yes, we should by all means, pursue the benefits of AI for our personnel and tools/equipment/weapons; but, never forget that a clever, sophisticated, sick and twisted adversary will likely find a way to undermine, and turn the AI against us. It is a tool; and, will provide tremendous benefits, but it should never be THE answer. As someone said long ago, “The greatest hinderance to progress isn’t ignorance; but, the illusion of knowledge.” RCP, fortunascorner.com