A few weeks ago, I recently came across a YouTube channel called ‘A Stable Genius’. This channel contains a small series of videos that really gives a good example of how sophisticated AI voice synthesis is becoming. The channel’s author explains, that they have developed and trained an AI on speech patterns of famous, real life, characters – in this case, Donald Trump & Barack Obama. While these videos are indeed lighthearted and intended for comedic effect, it really shows how far we have come in terms of being able to synthesize another person’s voice – amazing, but also very scary. Just imagine the havoc this could wreck in the hands of the wrong people – then we would really be talking about ‘fake news’. It seems whoever is in charge of the YouTube channel is very much aware of the potential implications of making this AI model publicly available.
It seems that there’s definitely a demand for making the model available, but also a lot of concern about how it might be used/misused. There were some great suggestions in the comments – which we are going to consider carefully. We’ll carry on working on the model behind the scenes, but won’t rush to release anything just yet.
Until that become a legitimate worry, enjoy the hilarity of A Stable Genius explaining various things.