Live data from Hacker News

AI Clones Your Voice After Listening for 5 Seconds (2018)

google.github.io

311–320 of 338 posts

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#311

Earlier quoted context omitted.

I told my wife that if I ever mention while on a phone call, she should know that I am in trouble an unable to speak freely. Sound like we'll all need more things like this eventually :(

how do you pronounce the symbols? i mean, 'redacted' is already a pretty strange thing to say by itself.

You just make static noises, like a radio signal being lost.

Also, you are lacking in the abstract thought department. Get that fixed, for your own benefit.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#312

Earlier quoted context omitted.

It would make sense to have another word that indicates it is genuinely you and you are genuinely speaking freely.

If you can speak freely you don't need an extra codeword to explain that you are using the codeword in it's real meaning. Unless maybe you suspect that somebody is listening to you and might learn your codeword from that.

The codeword is to make it clear your words should be taken 100% seriously without considering the risk you are being coerced / spoofed with AI. If I agree on a word in advance with someone that no one could possibly guess and insert into an attempt to coerce / spoof my voice, then if there is truly an emergency in which I need this person to wire money to a random account, they will actually do it because they will know my request is genuine.

If I'm being coerced, I could have a codeword to indicate that. If I'm being spoofed with AI, I'm not in control of "my" words, so I can't. I need instead to prove when I'm not being spoofed with AI. That's the purpose of this second codeword.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#313
post #311

Earlier quoted context omitted.

how do you pronounce the symbols? i mean, 'redacted' is already a pretty strange thing to say by itself.

You just make static noises, like a radio signal being lost. Also, you are lacking in the abstract thought department. Get that fixed, for your own benefit.

It's either a knowledge gap...or he is hopeless if he knew the symbol but didn't pick up on it. Nothing I know of can improve that.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#314

I once got a call while I was lecturing some students. It was repeated three times in three minutes - I assumed it was an emergency and stepped out. I was greeted by someone explaining that my father had caused a car accident, and they were calling on his behalf. That someone would need to send over some money for repairs or they’d call the police. Sure. They added that their cousin, the driver, is a parolee now hold…

How do you even prepare for something like that... Do we need to assign identifying keywords to each other when we leave home so we know we are really ourselves? Like a vocal pgp?

I'm broke so it makes things a lot easier.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#316

Earlier quoted context omitted.

I told my wife that if I ever mention while on a phone call, she should know that I am in trouble an unable to speak freely. Sound like we'll all need more things like this eventually :(

how do you pronounce the symbols? i mean, 'redacted' is already a pretty strange thing to say by itself.

They are pronounced "wakka" and "wakka" respectively.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#317

My bank's (HSBC) telephone banking offers the option do away with a PIN and instead a 'my voice is my password'phrase system. I'm glad I never opted in.

Oh god, Sneakers

passPORT

Also, that whole scenario would have been far less cool if they just recorded the dude for 5 seconds doing anything and pulled a whole CSI-style "put his voice into the visual basic GUI neural network and it works, bro!"

I love Sneakers.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#318
post #233

Earlier quoted context omitted.

I'm not really sure where you're getting this. It doesn't pick a specific voice from a database to use. From their introduction: "Our approach is to decouple speaker modeling from speech synthesis by independently training a speaker-discriminative embedding network that captures the space of speaker characteristics and training a high quality TTS model on a smaller dataset conditioned on the representation learned by…

They’re saying that underrepresented voices will have trouble being modeled. That matches my experience with this project: for example, I had a very tough time cloning female voices compared to nerdy-sounding / deep male voices.

It's more that the sounds produced during the recordings didn't cover the entire spectrum of possible sounds, so the model had to estimate their sound. All you really need is a paragraph which you can have the person read to get sufficient coverage or just enough recordings that it's not an issue anymore.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#319

Saw this on Minute papers last night and had a discussion with my partner about if we needed a secret password or not to tell if it were really one or the other on the phone. I figured that we had enough shared history that that wouldn't be a problem. Then we realized that there's no such thing as a simulated sense of humor yet and that that would be the best natural encryption to any communication.

2023: ai identifies your sense of humour after hearing you breathe for 0.7 milliseconds

Ha. There'd need to be so many more Ai breakthroughs to make that happen it'd be a little thing.

Re: AI Clones Your Voice After Listening for 5 Seconds (2018)

#320

Earlier quoted context omitted.

I think it would be cathartic to talk to someone you trusted but who is now gone. There's been decision points over the last few years where I would love to have just said my thoughts out loud to my dad and just have him nod and ask a couple of open ended questions so I could get it out. No specific guidance needed, just his particular style of listening. Clearly losing someone and being able to deal with it is an im…

"I think it would be cathartic to talk to someone you trusted but who is now gone." It would be cathartic, but in this case you wouldn't be talking to them but to a computer, who (at best) is pretending to be them. I think it's kind of creepy, when you really think about it, and it reminds me of the aversion the creator of Eliza had to his creation when he found out that people were spilling their guts out to it and…

Fair point
Post reply on HN