Following the documentary I watched yesterday about the influence of AI available here http://doyoutrustthiscomputer.org
The big problem I see about big corporates is that they don’t have any regulations to use their users data, and therefore can train their algorithms for free with a huge quantity of data, FB has more than 2 billion users which mean trillions of data shared to FB in a daily basis. You can watch the documentary I shared above to have a better idea of the danger those IA have and will have in our lives.
Big influencers start to stop using FB such as Tesla and SpaceX they just decided to remove their FB accounts with more than a million followers after the leak about the FB data used to influence the US elections.
I had a discussion with @geraldinestarke today about it and we came to the conclusion that the data is the most important part for those giants and if we found a way to give a control back to the users (aka FB users, Google users, twitter users…) we can limit the supra power dominance of those companies with regards to their AIs.
One way to go could be to introduce a user side API, where the company that wants to use their users data is enforced to comply to the rules of the user API and therefore the user is able to limit the usage of its data (posts, locations, images,…) we can even go further by introducing a token based crypto money where the company has to pay a rate (defined by the user) in order to have the right to use a limited quantity of data (or pay per usage), the user can at any moment stop providing its data to any company.
Technically speaking it is not impossible for a desktop browser we can have a firefox/chrome extension that takes care of that, in mobile it should also be possible with a background agent.
One use case could be with FB. The users that posts its data on FB will have its data instantaneously encrypted by those extensions and in order for someone else (another friend) to see your data it has to have an account on the same say platform that will limit its account to a certain quota. This quota has to be at a certain threshold that will reflect a single human behavior vs a IA that requires billion of data points to learn (we can think of 10000 requests per day). Beyond this limit the user/company has to pay per usage, only if the user allows him to do so. Each account in this platform is also verified and nominated to prevent creation of bots and multiplication of free quota.
Therefore if FB is willing to use your data and the data of million and billion of users that have chosen to use this platform they have to be granted by each user and they have to accept to pay a certain rate. The rate can be zero for those willing to give away their data.
This platform has to be open-sourced and preferably decentralized so everyone can gain trust into it by reviewing it’s code or auditing it, we can also improve its efficiency and its coverage with regards to the many data consuming softwares with the contributions of a community of developers.
Think of millions using it, new companies that are willing to create their own AIs will have the same power than giants since both have to comply to the same rules whatever their size.
If you are interested to help me build tbis solution or just pursue the discussions fell free to comment here. You can also reach me out by email: caner+user-api@candan.fr