Yes, I know, you’d rather watch a cat video, but this is important!
The video I’m featuring today details two legal developments – one in Germany, one in the US – that finally make the /people/ behind AI responsible for what their AI produces. And towards the very end, the video makes some very valuable points:
- There are ways of solving the AI inaccuracy problem but they haven’t been prioritised because ‘The market has rewarded speed and capability over accuracy and reliability.’ [https://www.youtube.com/watch?v=ehsq_0Cw6e4] Let me repeat that; the market has not rewarded accuracy and reliability. Why not?
- Because there were no negative consequences [to the companies] for the inaccuracies and lack of reliability.
- The legal system, globally, is starting to say ‘enough’. The negative consequences are starting to pile up.
- In the long term, forcing AI developers to make AI accurate and trustworthy will be of benefit to all humans because there is too much information for anyone, or even a group of anyones, to know it all.
- AI that can be trusted to synthesize this overwhelming mass of information could become the best tool humanity has ever invented. But only if it is accurate. Not 91% of the time, but all of the time.
This snippet is taken from the end of the video and is called ‘the silver lining’:
This next bit talks about how AI could be made accurate and reliable:
And finally, this is the whole video. It contains heaps of valuable explanations that provide context for the conclusion:
Given that the title and content of this post are about accuracy, I highly recommend that you check my conclusions to make sure I have not misrepresented what the video is saying. Close enough isn’t good enough.
Have a great weekend everyone,
Meeks
Meeka's Mind 