Joint AI Center (JAIC) acting director Nand Mulchandani said one of its first lethal AI projects is proceeding into a testing phase now. The Joint AI Center was founded in 2018 to act as the Pentagon’s leader in all things AI, and initially focused on non-lethal forms of AI. Mulchandani shared few specifics, but called the project “tactical edge AI” that will involve full human control and likened it to JAIC’s “flagship product” for joint warfighting operations.
“It is true that many of the product we work on will go into weapons systems. None of them right now are going to be autonomous weapon systems, we’re still governed by 3000.09,” he said. “The word lethality, and I think this is where the killer robot and Terminator stuff comes in, the edge case that everyone focuses on is such an outer edge case, and we are nowhere near it from a platforms, technology, capability, hardware, software, algorithms perspective to get anywhere close or near to that, but that obviously is where everybody jumps to.”
DoD Directive 3000.09 dictates how armed services branches of the military are allowed to automate weapons systems.
The Department of Defense put JAIC in charge of Pentagon AI initiatives last year. Mulchandani became acting JAIC director in May following the retirement of Air Force Lt. Gen. Jack Shanahan. In an interview in May, Mulchandani told VentureBeat said he worked for Silicon Valley startups before coming to the Pentagon a year ago.
Mulchandani and the JAIC discussed a wide range of topics today in a press conference including a “global war for AI talent” in the context of increasingly restrictive U.S. immigration policy, and asserted that the United States continues to lead in AI compared to China in many areas. Mulchandani said Chinese leadership in areas like facial recognition is due to human rights standards in the United States military. He said JAIC is not currently testing any form of facial recognition.
“It is not that the United States military is technologically incapable of developing such systems. It’s that our constitution and privacy laws protect the rights of U.S. citizens, and it’s how their data is collected and used, and therefore we simply don’t invest in building such universal surveillance and censorship systems,” he said.
Information Warfare is the newest of six JAIC mission initiatives. Use of AI techniques such as natural language processing (NLP) to condense and process information, including some supplied by private businesses, will play a role in helping along information warfare efforts, he said. NLP will also be at the center of Joint All Domain Command and Control systems, an initiative to help commanders triage communications.
More generally, Mulchandani talked about the “huge potential for using AI in offensive capabilities” in areas like cybersecurity. Mulchandani declined to offer any particular examples but generally pointed to examples like vulnerability discovery and weaknesses in networks. Lt Gen. Shanahan had also discussed issues of algorithmic warfare.
“You can read the news in terms of what our adversaries are doing out there, and you can imagine that there’s a lot of room for growth in that area,” he said.
Also detailed at the press conference were developments in recent months, including an $800 million contract with Booz Allen Hamilton, and Project Salus, a series of algorithms JAIC developed for NORTHCOM and National Guard units to predict supply chain resource challenges during the COVID-19 pandemic for things like water or available hotel rooms. JAIC will also be in charge of training military branches about AI ethics and putting the AI ethics principles adopted in February into practice.
JAIC is perhaps best known in tech circles as the creator of Project Maven, which led to protests and collective action by thousands of Google employees, eventually leading Google to pledge in its own AI ethics principles not to create autonomous weaponry.
Google and other tech giants continue to bid on federal contracts. A Tech Inquiry analysis released on Tuesday found strong ties between tech giants like Amazon, Google, and Microsoft, which hold thousands of lucrative contracts with the Pentagon and federal law enforcement agencies.
In other defense-related news, House and Senate committees recently approved amendments to the 2021 National Defense Authorization Act (NDAA) to include some guidance from the National Security Council on AI and set funding to help establish a national AI strategy.
Nand will join a panel talking about trends in edge computing and computer vision next week at VentureBeat’s digital conference Transform.
Credit: Source link