[mark_l_watson]

I liked his calling out the importance of commons for data. I especially like Common Crawl as a source of high quality web data, and I recently organized a meetup for Ocean Protocol which is a non-profit organization for sharing data.

Although Facebook and Google have some obvious problems with privacy, they also do a great service by open sourcing (with patent rights when using some of their open source deep learning projects) some tools that I really depend on for my work.

For the public commons, the public good, there is some strategy for leveraging open source and sharing data that respects privacy, and allows individuals and organizations to create valuable machine learning and AI applications. The are a lot of constraints: privacy, encouraging innovation, allowing fair profit from innovation without killing competition, etc.

In the past I found useful Lawrence Lessig‘s work on legal frameworks like the Creative Commons (I was the featured creative commoner for a few weeks, a long time ago; and, I have released all my recent books with a Creative Commons license even though I also sell copies). I think we need carefully thought out extensions to the Creative Commons licenses and ideas to cover data sharing to promote innovation and some room to earn a profit.