134
Mozilla Pushes Feds to Embrace Truly Open, Transparent AI Models
(www.pcmag.com)
This is a most excellent place for technology news and articles.
Above all else, the training data needs to be open. It's not really free software if the model is trained using a proprietary blend of training data. And it also opens the door for training data to be obtained unethically (which we know the AI companies are already doing)
No it doesn't.
Ethics are subjective.
The only people who say shit like that as a defence are people that 90% of people would agree are behaving unethically. Just because ethics overall are subjective doesn't mean there aren't some obvious right and wrong answers.