Using AI for digital resurrections and the potential legal ramifications

What would you say are ethical drawbacks of this tech that should be considered?

"I expect that, in the short term, the ethical concerns will be more prevalent than legal ones when considering the use of AI-powered grief-tech. Generative AI systems are becoming more powerful and life-like by the day, meaning users will be able to engage with the digital footprint of a deceased loved one like never before. It provides a real opportunity to provide meaningful interactions for grieving relatives, but runs the risk of preventing those relatives from recovering from a loss in the long term. There may also be risks in the future as the technology becomes more advanced and monetised: it seems more ethically concerning if a brand were to, for example, advertise through a digital recreation of a deceased loved one."

What major legal challenges do you foresee?

"Generally speaking, in the UK there are no “personality rights” per se, and data protection rights don’t extend to deceased persons. If the family of the loved one are those using the AI service, it’s unlikely that they will take issue with the use of their loved one’s content. However, if there is disagreement about the usage of a person’s online content in this manner, then the deceased person’s likeness may be less legally protectable than in respect of data relating to a living person.

The videos, photos or content of, or created by, a deceased person used to train the AI model will most likely have intellectual property rights attached to them. In some cases, where the photographer or videographer is still alive, they may well hold the copyright in this content. However, generally, it won’t be clear who might assert rights after death to the intellectual property rights in a person’s online presence - it is unlikely that someone will have thought of this in their estate planning. Families should consider the terms of any social media platforms that are being scraped to obtain this content as well. Big tech companies might have concerns with the reputational impact of AI grief tech being created from information derived from their platforms, and may retain rights if those platforms generated the content or avatar to begin with.

Finally, providers of the technology will need to take into account the developing legislative landscape, for example considering the new EU AI Act, and various protections under existing law, such as the additional protections afforded to children in the UK’s Online Safety Bill and the UK Information Commissioner’s Children’s Code. Because this is such a new space, legislators in different countries are likely to take different approaches, meaning that the protections available are likely to become increasingly complex."

An extract of Andrew’s comments were published in Newsweek, 30 July 2024.