Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

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

google.github.io

Post reply on HN
Print view View original

161–170 of 338 posts

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

#161

Post by hr5eaqhera »

BOBOTWINSTON wrote: post #73
pcmaffey wrote: post #37

Earlier quoted context omitted.

AI that detects AI seems entirely plausible. And like all “anti” measures, is another arms race (and if I put my scifi hat on, is what may lead to AI self-awareness).

Sort of reminds me of a talk Valve gave about creating an anti-cheat for Counter-Strike using AI. When asked if they were worried about people using AI to create cheats to fool the AI, his answer was essentially that it was an arms-race won by the person with more data/processing power. That person would most likely always be Valve. Link to talk: https://www.youtube.com/watch?v=ObhK8lUfIlc

It's a nice sentiment, but there are popular and easy-to-find cheating projects (not sure if I can name them here) that are still widely used, these projects have been active for years, before that talk was made, and still active today. Based on youtube videos and comments it seems many users are still using these cheats with little issue. And afaik, the one I'm referring to (initials P.I.) doesn't use any machine learning at all.

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

#162

Post by jdbernard »

nmeofthestate wrote: post #74

VCTK p240: duplicates the (a) north of England accent well. VCTK p260: all over the place accent-wise. LibriSpeech: can't really comment on the American examples, but they seem decent.

Sample 9 is a good example. The pronunciation of "biographer" is consistent even when it should be very different. All of the examples stress the first syllable but an American would stress the second.

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

#164

Post by aidenn0 »

ben_w wrote: post #154
nwsm wrote: post #119

Earlier quoted context omitted.

Images of us can be sourced from any number of places: social media, government surveillance, private surveillance. Video less so but from the same sources. Audio from phone companies, VoIP services, surveillance, etc. Health data easily from a number of private companies if you use new-age "health" services, or less easily (illegally) from health records. Maybe we can find solace in the fact that is or will soon be…

“Don’t worry, just about anyone can steal your soul and there’s nothing you can practically do to stop it” That doesn’t seem like a message of solace to me.

I am okay with the events that are unfolding currently.

http://gunshowcomic.com/648

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

#165

Post by bitL »

mistermann wrote: post #88
goodmachine wrote: post #7

The malign applications of this technology greatly outweigh the benign. Discuss.

Here's an outlandishly optimistic take on the possibilities.....as the internet and media is flooded with increasingly convincing but false representations of reality, a widespread habit of greater skepticism of "the facts" starts to spread throughout society, leading people to alter the speed at which they form new opinions on various issues (possibly degrading confidence in preexisting opinions), and the manner in…

Imagine we are actually a product of some advanced civilization bored with its capabilities where nothing seems real anymore, so they constructed us as a reality show to experience something that feels real.

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

#166

Post by agentultra »

Impressive but they still sound like robots imitating humans. I can only imagine the chaos this could cause, if used by bad actors, as it continues to improve. If someone took my voice I'm not sure that my partner would know it wasn't me. That would enable all kinds of social engineering attacks.

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

#167

Post by Enginerrrd »

agentultra wrote: post #166

Impressive but they still sound like robots imitating humans. I can only imagine the chaos this could cause, if used by bad actors, as it continues to improve. If someone took my voice I'm not sure that my partner would know it wasn't me. That would enable all kinds of social engineering attacks.

Exactly what I was thinking. We already have a problem with scammers claiming to be relatives who need some quick cash wired over. Imagine how much more effective that could be if it actually sound indistinguishable from a loved one.

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

#168

Post by therealfitz »

drbawb wrote: post #96
AbbasHaiderAli wrote: post #39

Wow, impressive results! Already a few examples in the comments of what bad actors could do this tech. I wanted to share an example of something good. I lost my dad about 6 years ago after a Stage 4 cancer diagnosis and a 3 month rapid diagnosis. I have some, but not a lot of video content of him from over the years. My mom still misses him terribly so for her 60th birthday I tried to splice together an audio message…

When I was doing a computer repair: I remember a woman coming in with a digital answering machine; the kind that stored its recordings in volatile flash. During a thunderstorm the night prior the machine lost power, and subsequently lost all the stored recordings. As it happens some of those lost recordings included messages from the woman's late mother. That moment has stuck with me for many, many years. The heartbr…

Thanks so much for sharing this story--this never even occurred to us when we created Google Takeout back in the day!

-Fitz

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

#170

Post by filoeleven »

Lordarminius wrote: post #12

AI can make decisions, create deep fakes, and now, clone voices. It may be that the next big business opportunity lies in creating 'anti-AI' technology just as it did with antiviruses in the 90's and 2000's

Better provenance is the way to guard against fakes.

For video and audio, I imagine a combination of hardware signing, perhaps with the camera itself living on an isolated, Secure Enclave-like chip, and sending hashes of (incoming images/video * deviceID * trustedTimestamp) to a blockchain or some other public distributed ledger. Getting the timestamp from a service that keeps its own record adds further security.

This obviously requires an internet connection, and would likely be useful mostly for news and government agencies, law enforcement. But if the culture is affected enough by deepfakes, I can imagine it becoming more ubiquitous. The parts are all there, it’s a question of utility.

Post reply on HN