A day or two ago I read an article about an AI sceptic who declared that AI is not going to be useful for much except "generating stuff", most of it textual and, while I share some of his scepticism with respect to the so-called intelligence with which practitioners and fans of the tools credit them, I tend to disagree with him about its capabilities.
I have mentioned elsewhere that we are the reason AI will kill people and how it will be done and while that is not the topic of this 'discussion', it is important to know one underdeveloped capability of AI systems which can be leveraged to achieve that end.
One of the most exciting and prominent developments to come out of the field is the phenomenon they call "deep fakes", in which AI is made to make still images of people look like they are moving. This has been used to great effect to develop caricatures of many people saying and singing absurd things, but I have not yet seen it applied to entire human bodies.
What I have seen applied to entire human bodies is generative AI, also to great effect, and with increasing levels of realism which has made some white people cry about its dangers. Sometimes I think these people are just being jealous.
So, what if we combine both and see what comes out?
While deep fakes are in a bit of a shallow place right now, I know how they can be enhanced in terms of power and accuracy.
How? Through the use of motion capture and 3-D imaging techniques.
I personally prefer motion capture because there is a lot of raw data in the hands of different individuals and companies, the technology is mature and, strangely enough, motion capture is easier to map to generated images than 3-D imaging techniques like lidar and wi-fi.
Also, I am quite sure motion capture data is a lot less complex to analyse and classify than 3-D imaging probably because it lacks its granularity or resolution and also because motion capture data is a lot less than 3-D imaging data, which reduces the number of strong computers to use when processing the data.
As a result, entire fully AI-generated videos are in the future because, as usual, I know someone is already in some basement somewhere scratching his balls coaxing miracles out of several lumps of sand. Viva amigo!
Personally, I am looking forward to the day when entire movies will be made using AI. The prompts are already there and the films coming out of the 'woods worldwide could easily be made using those templates - I have talked about them elsewhere.
At that point, human creativity and imagination will not be limited by the tools available, but by people's willingness to say exactly what is in their hearts.
I know one type of entity that will use this capability once developed to achieve their ends: governments. With their desire for increased power and control over people's hearts and minds, governments across the planet will use these tools to kill a lot of people. Already, China is suspected of sending suicidal propaganda to Tiktok users in the West - and this with just recommendation algorithms. Soko imechafuka.
HA! None.