Want to chat with Shakespeare? AI bots will soon allow us to talk to the dead

“Sorry, the line is a little fuzzy—can you speak up?”
“Sorry, the line is a little fuzzy—can you speak up?”
Image: Reuters/Cheryl Ravelo
We may earn a commission from links on this page.

Imagine debating the interpretation of a Shakespearean sonnet and being able to clarify its meaning with the bard himself. Or sitting in history class and being able to ask George Washington questions about the Constitution, no soul-conjuring witchcraft required.

In the next decade, advancing AI technology will allow us to learn from the dead first-hand. New chatbot programs are being developed to keep our knowledge active after our physical being passes away.

Early research in this topic already allows us to simulate dialogues with the dead. For example, Russian startup Luka has created a simulacrum of the notoriously private musician Prince. This AI-powered chatbot draws from song lyrics and rare interview snippets to let users instant message with a vision of the late singer, who died in 2016.

But tech necromancy is only just getting started, and new developments will soon make afterlife conversations commonplace. Dr. Hossein Rahnama of Ryerson University and the MIT Media Lab is working on a digital afterlife technology he’s termed “augmented eternity.” To create an “immortalized” augmented-eternity bot, data produced by a living individual—such as personal emails, tweets, text messages, and any published works—is fed into an artificial neutral network. This model brain is then able to process and replicate stylistic nuances of speech and complex patterns of thought to mimic natural expression.

But augmented eternity aims to do more than memorialize an individual: It will harness the educational potential of immortality, too.

“Augmented eternity can be used as an educational tool that allows people in the future to ‘speak’ with historical figures,” Rahnama explains. In this way, a brilliant mind could be digitally preserved and used as an academic resource for generations to come. AI tutor bots like virtual math coach Whizz Education and exam-prep bot Gojimo are already proving to be powerful tools in increasingly overstuffed classrooms. But augmented eternity promises be unique in its ability to help students deepen their understanding of an actual human being, not a manufactured AI bot.

The best augmented-eternity bots will be created in the image of prolific people who have created a significant amount of collectable data. Because of the amount of text-based data they produce, influential writers will be especially well represented as augmented-reality bots, which will most benefit academics studying English or the humanities. Geniuses in the math and sciences who produced little in the way of a written record may therefore be harder to resurrect.

The further back in time we go, the harder it will generally be to create an accurate depiction of an individual. This is because we create so much more traceable data in contemporary times; hand-written notes and physical books from pre-internet subjects would have to be digitized in order to be programmable. Younger generations generating high volumes of collectable information will therefore be plum picks for more convincing resurrections, and Rahnama notes that many modern thinkers will now leave behind approximately a zetabyte’s worth of text.

Rahnama’s vision for augmented eternity’s educational application focuses on “swappable identities,” where the same question can be addressed to AI personas with drastically different backgrounds. Being able to directly speak to different primary sources on historical issues could be an invaluable, perspective-enhancing tool for students. “The future is about being able to switch your lens and see the world from someone else’s view,” he says. “Issues such as gun control, liberalism, genetic cloning, and legal disputes can all be seen from different political, scientific, academic, and statistical angles.”

Augmented-eternity bots are designed to continue applying an individual’s old outlook to new information. This means they would not be limited to the knowledge accrued in their lifetimes, but rather fluidly stay abreast of current events and discoveries. Therefore, students could consult leading thinkers about events that occurred years after their deaths. For example, when asked about US president Donald Trump’s views on job creation, “activating Ronald Regan’s persona would probably generate a different answer in comparison with activating Jimmy Carter’s persona.”

One of augmented eternity’s key hurdles is privacy. How to maintain an individual’s privacy while allowing students to pick their digitized brain is still being questioned. Interrogating an augmented eternity persona about intimate aspects of their former humanity—their feelings for their loved ones, or their most embarrassing secrets, for instance—would be inappropriate and morally transgressive. This will likely be overcome by the implementation of controlled, subject-specific chat sessions that restrict access to different memory recalls.

It’s hard to say how the ability to chat with future Shakespeare and Washington equivalents will affect society’s relationship with historical figures. Yet as history repeats itself and humans prove forgetful of lessons mired in the past, immortal advisors may become invaluable purveyors of life lessons unavailable to the living. How better to understand the past than from someone who was there?