Most of the most powerful machine learning techniques require very large amounts of carefully curated data.
This requirement for data gives powerful economies of scale to the Big Tech Companies, since they have many of the biggest repositories. The data in their possession is therefore jealously guarded.
However, an emerging technique, Homomorphic Encryption, promises to enable one to have one's cake and eat it too. As an analogy, HE enables one to share the 'shadow cast by the data', but not the data itself. Meaning that one can 'recognise an object', but not the most personal specifics.
It also enables one to rescind access to data after the fact, a perfect form of digital rights management, which can create a measure of safety in a world where vast databases are regularly pwned, or carelessly exposed on lost flash drives.
Anonymous data can be anonymously traded and leveraged, meaning that one can arbitrage one's own datasets without the risk of compromising one's privacy or identity.
These same techniques combine very nicely with smart-contract based organisations (DAOs and similar), such as this new form of gamified distributed financial instrument.
HE enables truly infinite games, since game pieces cannot be removed from play.
