Rapes and murders committed in virtual reality in the Metaverse might have to be treated as criminal offences because of the real life impact on the victims, the head of the National Crime Agency has said.
Graeme Biggar, the NCA’s director general, said that people would “feel a physical manifestation” while wearing a haptic suit - which allows a wearer to sense actions carried out in virtual reality - if targeted by a sexual or other violent attack.
He said this was something that law enforcers “need to prepare for” and that his agency was already engaged in trying to work out how to police the Metaverse.
His comments came in an interview with the Evening Standard in which he set out the challenges and opportunities presented by rapidly developing artificial intelligence technology.
He said the benefits could include helping law enforcement break the encryption used by criminals to communicate in private as well as making it easier to analyse huge data sets to stop criminal networks.
The dangers included allowing a potential increase in fraud as criminals used AI to compose more convincing scam emails and texts and a similarly improved ability for paedophiles to groom children.
But his most striking remarks were on the potential for crimes to be committed in virtual reality on the Metaverse.
“We are beginning to think about what is a crime in the Metaverse and how we police it,” Mr Biggar said.
“It’s not dominating our thinking because there is plenty of real world crime for us to be getting on with, but if you are in the Metaverse wearing a haptic suit where you can sense what is happening to and then you are sexually assaulted, raped or murdered, even if you are not wearing a haptic suit is that ok?
“If you are wearing a haptic suit then you are going to feel a physical manifestation of that rape or murder. All of those, there’s the beginnings of them now but the potential for them in the future and that is part of what we need to prepare for, to help people reap the benefits of exciting technological developments, while still staying safe.”
Mr Biggar’s remarks will add to the debate about how best to regulate AI to maximise its benefits while protecting the public from its potentially malign effects.
He said the “threats are fraud and cyber crime” including the “real risk that AI will be able to copy your voice pattern and fool” the voice recognition used by banks to verify the identity of callers.
He said that banks and the NCA were preparing to combat this but that a further threat was presented by the ability of artificial intelligence to generate more convincing text.
Warning that there is already “an awful lot of online fraud” Mr Biggar said “there will be even more if all the emails and phone calls and texts you got were written very convincingly or had better scripts.
“I’m sure that AI and ChatGPT and other brands that are available will be used to write more professional scripts that will make more people more likely to fall for them.”
Mr Biggar said that AI quantum computing might also be able to “defeat the encryption that we or banks are using” in the future, while a further threat was “the ability to generate deep fakes of children mixed with images of real children to create a different type of indecent image of a child.”
“We’ve seen hints of that already but it could happen more,” he said. “The ability to merge in a realistic way images of children that are on the internet, real kids, with other images automatically generated of them being abused.
“Paedophiles could get better scripts to groom children. Young girls and boys have been extorted, blackmailed into exposing themselves, doing things to themselves, self-harming themselves, because someone they are in contact with online gets something from them and blackmails them into doing more and more and sometimes at scale. If they have got a better script and ability to do that then that will be more damaging.”
Mr Biggar said law enforcers were preparing all the challenges and would benefit from some of the other capabilities of AI.
One example was the ability of machine learning to analyse bulk data and spot criminals’ connections and networks more effectively. The potential use of AI to defeat the encryption used by criminals to communicate presented a further opportunity.
