AI and Posthumous Digital Personas

Kazuo Ishiguro’s newest book, Klara and the Sun, is a work of fiction, but barely so. Klara is an AF, or artificial friend. Spoiler alert — her purpose is to study her owner so intensely that she can not only learn her habits and perform her activities but take on personality characteristics to the point where she could act as a post-death extension of her owner. Preposterous as this may sound, it’s not so far from the current truth of artificial intelligence (AI).

As reported by the Wall Street Journal,

Researchers and entrepreneurs are starting to ponder how artificial intelligence could create versions of people after their deaths — not only as static replicas for the benefit of their loved ones but as evolving digital entities that may steer companies or influence world events. “Could AI Keep People ‘Alive’ After Death?”

Wall Street Journal, July 3, 2021

Tech companies are already working on digital personas, training them with AI to interact with humans. Microsoft has a patent, not yet publicly exploited, on a conversational “chatbot” that can take on the characteristics of a historical figure or — like Klara — a family member. The chatbot is trained using social media posts and text messages of the deceased. A person could even create his or her own digital persona to retain posthumous control of a business. (See Dalvin Brown, “Innovations,” Washington Post, February 4, 2021.)

The legal questions are dizzying, and the ethical ones even more so. Who owns the digital persona? Who is entitled to its earnings? Who owns the copyright to works it creates? Who can enforce the rights of the deceased in texts and social media posts? What privacy and, where applicable, publicity rights will put the reins on the exploitation of the technology? What are the emotional implications of using technology to address grief?

We’ve previously referenced the use of AI to bring historical characters “back to life” for educational purposes and we have explored the use and limitations of copyright, publicity/privacy and defamation law to combat the spread of disinformation by “deepfakes.” It is dubious that our existing intellectual property system will provide long-term answers to the myriad of questions this use of AI will raise. Legislation at the state or preferably federal level may soon be needed to control the new world that Ishiguro has imagined. But, speaking of imagination, it’s hard to see how consensus could be achieved around the complex moral issues involved.