THE government’s pledges to act “responsibly” in developing AI weapons have fallen short of reality, a Lords committee has found.

In a report by the AI weapons systems committee, published today, it acknowledged that autonomous weapon systems (AWS) — known as “killer robots” — were “one of the most controversial issues of AI today.”

However the government’s aspirations to be “ambitious, safe, responsible” have “not lived up to reality,” it said.

Among the committee’s concerns was that the Ministry of Defence had resisted giving an operational definition of AWS. The panel argue however that it is necessary to aid the creation of meaningful policy and involvement in international discussions.

Secrecy has also shrouded elements of policy development, with the MoD refusing to publish minutes and reports prepared for the panel.

“The fast pace of development, as well as the lack of publicly available information on how AI is being developed, pose issues for parliamentary scrutiny,” the report says.

Calling for greater transparency, the committee underlined that “arguments of secrecy must not be used to sidestep accountability.” 

Peter Burt, researcher at campaign group Drone Wars UK, said that the government is “racing ahead to develop autonomous weapons regardless of the risks they pose.  

“Consideration of the ethical hazards and harms from autonomous weapons has lagged far behind research into the weapons themselves.”

Mr Burt said it was “telling” that the high-profile AI Safety Summit ruled out discussing military applications of AI, which are “the most harmful potential uses imaginable.”

“As the Lords have concluded, a key goal is accelerating efforts to achieving an effective international instrument to control the military use of AI,” he said.

“If ‘Global Britain’ is to have any meaning the UK government must now throw its weight behind the growing movement to ban fully autonomous weapon systems.”

CND General Secretary Kate Hudson said there was “no ‘responsible’ or ethical way” to use AI militarily.

“AI technology is advancing rapidly and we urgently need domestic and international legal frameworks in place to regulate its use in order to protect people”, she said.