In the very near future, we could see artificial intelligence (A.I.) as an integral part of the military, because it’s already being used by the military today. Recently, the U.S. Government has been striking deals with A.I. companies to deploy the technology militarily. Furthermore, the ongoing military operations in Iran have no doubt increased the urgency for the government to acquire this technology. However, some A.I. companies have been pushing back on the use of A.I. for military purposes, such as piloting autonomous weapons.
The Significance of the Anthropic Lawsuit
Anthropic is an A.I. company that has been developing an LLM (large language model) called Claude, which the government wants to integrate into the military. The deal between Anthropic and the government fell into conflict when Anthropic made it clear that the A.I. came with two major stipulations: It couldn’t be used for “domestic mass surveillance” or “fully autonomous lethal weapons”. The conflict was further escalated by the Pentagon’s response of retaliation when they blacklisted the company and labeled the company as “a supply-chain risk”, which refers to a company as having the potential to intentionally disrupt and or sabotage systems of the economy. The government argued, however, that the restrictions Anthropic imposed would interfere with its ability to conduct military operations using A.I. and undermine its effectiveness. To counter this, Anthropic decided to sue the government for violating their First Amendment rights with this blacklist and labeling, and the situation is still developing. Meanwhile, amid this debacle between Anthropic and the Government, OpenAI agreed to a deal with the U.S. military.
OpenAI’s Deal with the Government
OpenAI took advantage of the situation and secured a deal with the government; however, the company faced significant backlash for not including the similar language Anthropic had, which restricted the government from mass surveillance and the use of A.I. to autonomously pilot weapons. Due to this, OpenAI responded by changing the language included in their deal with the government to more specifically ensure that the A.I. wouldn’t be used in ways that violate their redlines which were, “No use of OpenAI technology for mass domestic surveillance.” as well as, “No use of OpenAI technology to direct autonomous weapons systems.” and lastly “No use of OpenAI technology for high-stakes automated decisions (e.g. systems such as “social credit”)”. The company with these changes claimed they had more “guardrails than any previous agreement for classified A.I. deployments, including Anthropic’s,” which, while possibly true, some believe is more about P.R. recovery from the all controversy this deal had stirred up, rather than truly an attempt to restrict the government out of ethical concern. However, all the discussion about the capabilities and limitations of A.I.’s implementation in the government and military raises the question of what A.I. could be capable of when it is acquired.
What are AI’s Capabilities?
The acquisition of the A.I. could lead to several possibilities for its usage in the military and government. According to MIT’s Technology Review, one possible use could be for the military to use A.I. to “rank lists of targets” and recommend which to “strike first,” based on a statement from a Defense Department official. Additionally, the A.I. could review and analyze “aerial drone footage” to assist in deciphering potential targets. Furthermore, with the current situation in Iran, some news outlets have called into question whether A.I. has already been involved with the bombings and the locations chosen by the U.S. military. Earlier last month, Claude A.I. had already been involved in the operation in which Nicolas Maduro was caught and apprehended by the U.S. in Venezuela, though the A.I.’s role in the operation remains unclear.
During this time of conflict, it’s likely that we will see conversations about the use of A.I. in government and the military continue, especially regarding the ethical application of such tools. Ultimately, the lawsuit involving Anthropic will help determine the future of A.I. and the rules and regulations revolving around it. Given the strong possibility of future government use of A.I., people currently and those approaching adulthood should express their views to lawmakers, as legislation is being drafted.
