For decades, people have been able to improve the quality of pictures using image editing software like Adobe Photoshop.
Now, Adobe has just demonstrated a new tool they are working on, which could have an even bigger impact.
It’s called VoCo, and it analyses people’s speech soundwaves to make it easy to change what a person said.
Check out the video above for a quite terrifying demonstration.
The software works by analysing the waveform of the sound of someone speaking, and using voice recognition to match the sounds to certain words.
Then, the software can essentially be used to completely change what a person originally said, as simply as changing the text in a word processor.
In the demonstration above, the system could change the order of words or delete certain words from the presenter’s speech just by selecting and deleting the words in a window.
However, the truly mesmerising part came when the system was used to actually change what was said.
Since the software can analyse and match how a person says certain words (including their pitch, speed and potentially even accent), the next demonstration showed the user typing completely new words into the system, which were then instantly said out loud in the original voice!
Just imagine how this could affect a number of creative industries, from film to music to journalism, and above all justice and law enforcement.
Deceased singers like Elvis and Michael Jackson could be made to “sing” new songs in their original voice.
Films could edit dialogue in post-production without the actors needing to re-shoot scenes.
“Evidence” of phone conversations with politicians could be altered to make it appear they are saying things which never happened.
Fortunately, Adobe says it is working hard to implement a digital watermark into any audio, to ensure that it will be easy to confirm whether the things being said were “real” or not.
Either way, the next time I get a phone call from Winston Churchill next year, I’m going to be even more skeptical.