Okta’s got an interesting idea up its sleeve: why not ditch those pesky passwords and let our voices do the talking?
This identity management champ is cooking up a patent for a multi-factor authentication system. But here’s the twist: instead of just asking for your password or sending a text to your phone, it wants to use AI to analyze your voice. Picture it: you say a few words, and the system thinks, “Yep, that’s definitely Sarah!” But don’t toss out your passwords just yet. Okta’s voice feature is just one of several authentication methods they’re considering.
How does it work? They’ve trained a neural network to identify users based on just a few spoken words. And the cool part? It doesn’t need to match those words to any specific text. So, even if you’re singing your heart out or whispering a secret, it’ll recognize you. It’s all about catching the unique bits of your voice: the pitch, the way you pronounce things, even the acoustics.
Voice recognition isn’t just Okta’s brainchild, though. Big names like Meta and PayPal are also getting into the “voiceprint” game, where the distinct sound and rhythm of your voice is your ID card.
Ari Weil from cybersecurity firm Cyera broke down why this method’s gaining traction: it’s just simpler and friendlier for users. Plus, organizations trust it more for tackling fraud. But, and here’s the kicker, with AI making those eerie realistic deepfakes, there’s a twist in the tale. Research from University College London highlighted that most folks could only spot deepfaked voices 73% of the time. As Weil put it, “With AI getting so good, verifying through biometrics is crucial.”
But what about those deepfakes? Okta’s on it. They’re planning to store a user’s unique voice data and always cross-check any future audio snippets against it to ensure they match up.
Yet, there’s a cautionary note from Aubrey Turner of Ping Identity. Biometrics, for all their uniqueness, have a chink in their armor. If someone snags your biometric data, you can’t change it like a password. And with AI’s double-edged sword, while it boosts security, it’s also empowering cyber baddies with fancier phishing tactics and, yep, more deepfakes. And considering Okta’s past hiccups with data breaches, this could be a concern for some.
