Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News Editorials & Other Articles General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

General Discussion

AZJonnie

(3,915 posts)

3. Probably can replicate the tone with a 3 second sample, but there's no way it would fool someone who knows you well

Reply to doublethink (Original post)

Tue Dec 23, 2025, 05:00 AM

Dec 2025

It would need a much larger sample to hear the ways you inflect words in conversation, which is every bit an important as the frequency/tone. And then there's speaking style, including things like your use of vocabulary. In 3 seconds it can't determine if you speak at the level of someone barely literate, or a Ph.D.

So, yeah, it's a little scary esp. if used to misrepresent you publicly, but I'm a lot more worried what it could do with 3 MINUTES of you talking, than what it can with 3 seconds, put it like that.

Edit history

Please sign in to view edit histories.

Recommendations

3 members have recommended this reply (displayed in chronological order):

raccoon Orrex doublethink

4 replies

= new reply since forum marked as read

Highlight:

A.I. only needs 3 seconds to hear your voice and replicate it. [View all] doublethink Dec 2025 OP

"Ah ee ooh" unblock Dec 2025 #1

A few years back, I read that voice recognition could be used for security protocols. John1956PA Dec 2025 #2

Probably can replicate the tone with a 3 second sample, but there's no way it would fool someone who knows you well AZJonnie Dec 2025 #3

And what if it's somebody legit calling you back about something about your medical care? raccoon Dec 2025 #4