The Download: striking actors training AI, and breaking ‘unbreakable’ encryption

Between July and September this year, actors in the US were invited to participate in an unusual research project, designed to capture their voices, faces, movements, and expressions.

The project, which coincided with Hollywood’s historic strikes by the Writers Guild of America and the Screen Actors Guild, was run by London-based emotion AI company Realeyes and Meta. The information captured from the actors was fed into an AI database to better understand and express human emotions. 

Many actors across the industry worry that AI could be used to replace them, whether or not their exact faces are copied. And in this case, by providing the facial expressions that will teach AI to appear more human, study participants may in fact have been the ones inadvertently training their own potential replacements. Read the full story.

—Eileen Guo

Inside the quest for unbreakable encryption

When we check our email inbox, log in to our bank accounts, or message on Signal, our passwords and credentials are protected through encryption, a locking scheme that uses secrets to disguise our data. 

Our trust in online security is rooted in mathematics. Encryption schemes are built on families of math problems called one-way functions—calculations that are easy to carry out in one direction but almost impossible to solve efficiently from the other, even with a powerful computer. 

There’s a problem, however. Although mathematicians suspect true one-way functions exist, they’ve yet to prove it. This conundrum haunts all encryption. Our data is secured by the fact that no one knows how to crack the schemes that protect it—or at least not yet. Read the full story.