Legislation call to prevent AI ‘ghostbots’

A study conducted by researchers at Queen’s University Belfast (QUB), Newcastle University Law School, and Aston Law School says that increased understanding of ‘ghostbots’ and AI-specific clauses in wills could stop humans from being digitally reincarnated without our consent.

‘Ghostbots’ describes the phenomenon of using AI to convert data – often collected from social media platforms – to create digital renderings of the dead. These can include replicas, holographs, chatbots, or deepfakes that mirror the speech patterns, movement, and bevhaiour of people who have passed away.

Recent celebrity ‘ghostbots’ have seen a hologram of Tupac Shakur perform at Coachella, Queen guitarist Brian May duet with a hologram of his late bandmate Freddie Mercury in 2022, and Kim Kardashian receiving a 40th birthday message from a deepfake of her late father seventeen years after his death.

The study’s authors are keen to flag the potential legal, social and reputational ramifications of such avatars.

Dr Marisa McVey, from the School of Law at QUB, says that while ghostbots represent an intersection of various areas of the law, the lack of UK privacy and data protection laws constitutes a disturbing lack of consideration for the deceased’s post-mortem dignity: “While it is not thought that ‘ghostbots’ could cause physical harm, the likelihood is that they could cause emotional distress and economic harm, particularly impacting upon the deceased’s loved ones and heirs.

“Currently, in the absence of specific legislation in the UK and further afield, it’s unclear who might have the power to bring back our digital persona after we die.”

Recent legislation changes in the EU and US have underlined the potential for further regulation of ghostbots.

In the state of New York, laws have been introduced prohibiting the digital replication of deceased performers, and celebrities have a right of publicity to control the commercial use of their name and image, which endures for forty years following their death.

In Europe, the proposed EU AI Act would also require AI-generated disinformation to be clearly labelled.

The paper’s authors have suggested that, in the absence of current legislation, UK citizens could protect their post-mortem privacy by including legally binding ‘Do not bot me’ clauses in their wills: “We also suggest that in addition to legal protections, greater societal awareness of the phenomenon of ‘ghostbots’, education on digital legacies and cohesive protection across different jurisdictions is crucial to ensure that this does not happen without our permission.”

The paper, ‘Governing Ghostbots’, is part of the Leverhulme Trust-funded project ‘Modern Technologies, Privacy Law and the Dead’.

Related: Industry leaders brand AI advances ‘dangerous’

Related: FDA green-lights human trials of brain implant technology