Certainly, AI can get predictions wrong. From what I know, the idea of AI is to make corrections and get better. Machine Learning and Deep Learning are example. DARPA is actually using the latest tech to come up with an AI interface that can respond to user questions. The DARPA project would be able to explain HOW it came to a decision. Thus, if an AI does get something wrong, then a user will know precisely where to correct.
NOIA lists AI as part of its CSL within the technical paper:
There are two models of how Content Scaling Layer Organisations can be implemented on the NOIA Network: ● CSL Cloud Controller Model – a single centralised point in NOIA Network which controls other CSL system layers ● CSL DAO Concept Model – a fully decentralised organisation with self imposed governance
Consider this two model approach with the eventual full decentralization, and NOIA can use AI in alpha to develop a fully viable solution in Beta progressing as it believes its ready with minimal error (or if AI tech advances to the point that corrections become easy).
Any additional thoughts?
Thanks For the clarification @machnbirdsparo
If i maY: recently two groups in the telegram channel who use the bot for Their Quiz Had One similar issue "The bot send questions with no options and also send questions with two answers "
Not to talk of virus making the Bot go Wired andd start t sending repeat information or not send you information at all as in the case of ginabot some Time AgO.
Even when the error was corrected but damage as already be made.
Nice writeups @machnbirdsparo
I believe @noia has great potential
I am not sure I understand your illustration @ogt. Though, I di get that the telegram botwas unable to appropriately respond to human users.
The DARPA AI project is specifically design to query the AI program itself so a human can understand how the AI program operated. Mind you, DARPA's AI is a hefty project. Last I checked, they had not solved the problem with much significance.
Anyway thanks @ogt. Feel free to continue this conversation.
Thanks @machnbirdsparo
For a great post, im clarified