The US Defense Advanced Research Projects Agency (DARPA), the Pentagon's research unit, has announced plans to invest up to $2bn in AI weapons systems over the next five years.
The agency claimed its latest campaign of investments would aim to accelerate the "third wave" of technological innovations, which AI technologies is at the heart of.
In a statement released on the agency's website, DARPA explained that the new program, called the "AI Next" campaign, will look to "address the limitations of the first and second wave AI technologies". This next wave of tech will focus more on "contextual adaptations" which will lead to more collaborative and trusting "partnerships between humans and machines".
Visit Innovation Enterprise's DATAx New York on December 12–13, 2018
AI-enabled weaponry is a controversial topic as the notion of algorithms and AI determining and executing decisions of life and death is an unpopular one with American voters. In June 2018, 4,000 Google employees signed a letter to Sundar Pichai, CEO of Google, petitioning him to exit the contract called "Project Maven". The project saw Google using AI systems to help analyze archives of drone footage taken over Iraq and Syria. Despite Google claiming it would not be directly involved in flying or launching any weapons, Google chose not to renew the contract amid pressures from staff.
The project appeared to be more focused on building up the partnership between humans and machines, with less of a focus on previous sentiments by the Department of Defense (DoD), such as when it claimed the US would be "winning wars with computer algorithms". In a statement made at the DARPA D60 Symposium, agency director Dr. Steven Walker elaborated the agency's goals with this new campaign.
"With AI Next, we are making multiple research investments aimed at transforming computers from specialized tools to partners in problem-solving," Walker remarked. "Today, machines lack contextual reasoning capabilities, and their training must cover every eventuality, which is not only costly, but ultimately impossible.
"We want to explore how machines can acquire human-like communication and reasoning capabilities, with the ability to recognize new situations and environments and adapt to them."