Know

AI Can Now Bring Back the Dead, but Is It Ethical?

3 min read
Jan 28, 2021
  • Post on Twitter
  • Share on Facebook
  • Post on LinkedIn
  • Post on Reddit
  • Copy link to clipboard
    Link copied to clipboard

South Korean national television broadcaster SPS will showcase a performance this week with new songs from Kim Kwang-seok, a famous South Korean singer. There’s just one catch: Kwang-seok has been dead for 25 years.

Likewise, Microsoft received a patent in December that allows for creating a conversational chatbot that is based on a “past or present entity.” In other words, a chatbot modeled after a specific person, living or dead, that you can have a conversation with. The patent application was filed in 2017.

Recent developments in artificial intelligence (AI) can, in a sense, bring back the dead. It’s something that sounds straight out of an episode of “The Twilight Zone” or “Black Mirror.” Oh, wait. In the case of “Black Mirror,” that actually happened.

Both developments have come along as part of the latest evolution of AI.

Kwang-seok’s performance will be part of a program called “Competition of the Century: AI vs Human." Instead of competing against a human competitor, though, Kwang-seok will be performing a duet with a human. 

This isn’t the first time South Korea has used AI to resurrect a famous performer. On New Year’s Eve, BTS had an online performance with an AI version of its late singer Shin-Hae-chul. Before that, in December, Mnet, a music channel, broadcast a show titled “One More Time.” The program paid tribute to late artists by showcasing their work with AI and holograms.

In Microsoft’s case, the potential product would use social media posts, images, messages, and other “social data” from the person being recreated. The chatbot would then be trained with this data to have a conversation. If a user were to ask the chatbot a question that couldn’t be answered with that data, other sources of data could be used.

However, Microsoft has no plans to create a product based on the patent and the technology described in it. In the almost three years since the patent was filed, Microsoft has formed several AI ethics groups that review its products and inventions.

Besides being creepy, it’s not hard to see the ethical issues that can arise, especially since the technology can build recreations of real people and not just fake ones. For example, if new works or songs are created, who owns them? The family of the deceased person, the creator of the AI, or even the AI itself? The act of replicating someone’s voice also carries the risk of fraud, misinformation campaigns, and contributing to potentially fake news.

For the evolution of AI to continue, these dilemmas, among others, will need to be addressed. Around the world, the call for legal guidelines and guidance on ethical issues for the use of AI is reaching a crescendo. That will only continue as the technology continues to be developed. Those guidelines will need to provide regulations that protect society while still allowing for AI technology to evolve.

Sources