This could use some explanation: > "Mozilla.ai’s initial focus? Tools that make generative AI safer and more transparent. And, people-centric recommendation sy

This could use some explanation: > "Mozilla.ai’s initial focus? Tools that make generative AI safer and more transparent. And, people-centric recommendation systems that don’t misinform or undermine our well-being." 1) What kind of tool is that? More 'transparent' - are we talking about a map of the training corpus, snapshots of the AI's internal state, or what? And what does 'safer' mean - a filter on the output of some kind? Will that be a transparent filter? Will there be a secretive committee that decides what's safe and what's not? Will it take recommendations from the Department of Homeland Security on that question, as Twitter was doing? 2) What is a people-centric recommendation system, exactly? Does that mean people get control of the knobs and dials on the recommendation system? For example, I'd like an 'inverted' option, the ability to grab a random selection of the content that the system thinks I wouldn't care about. I'd like the option to make different lists of channels and randomly grab content just from those lists. Also, a list of channels I never want to see any content from, period. I suspect, however, that the result will be more in the flavor of "Big Brother knows what's good for you, and what's not". As far as whether I'd be 'misinformed' or 'well-being destabilized' by AI output, please stop being so paternalistic. If you're afraid people can't distinguish between fact and fiction, invest in better public education and teach skepticism and critical analysis skills to young people, instead of training them to be obedient little zombies who have absolute faith in institutional authorities.