Data as "Food" for AI and Why We Need Data Sovereignty SOON

There is an ongoing question floating around about when our technology will begin eating us. An argument can be made that it is already happening. Imagine for example that many companies which form exist merely to feed artificial intelligence. Imagine these companies are social networking sites, or even blockchain technologies like Steem which embrace "transparency" of the data with the purpose behind it being to feed the AI.

On Steem in theory it is supposed to be a mutualism. You contribute your data but you can earn something from the reward poll in exchange for it. In other words the blockchain might very well be "food" for an AI but it's also capable of feeding you. In my opinion this is very similar to the relationship between humans and gut micro-organisms where we have a mutualism. The gut micro-organisms supports the complex adaptive ecosystem which is our body by helping us to digest our food, allowing us to get what we need from our food whilst at the same time they get what they want from the process of digesting our food.

In the ideal world, where data is "fuel" to machines, or "food" to AI, we maintain ownership of our data. This ownership however is impossible to have if we have no mechanism to control the data we share. Data sovereignty technology is a means of maintaining control over data as we share it with the machines. The Secret Contract technology of the Enigma Project is a perfect example of a data sovereignty technology. Data Wallet is a project which leverages the Enigma Protocol to provide the fundamental building block for data sovereignty.

Data sovereignty is an ethical stance. Not every company, or every project, is willing to take the stance that you own yourself, that you own your mind, that you own your data. Some companies believe that your eyeball are a resource to be captured and spent. The companies which have these ethics are also capable of building AI which must eat your data to grow. If you do not have privacy then you cannot maintain data sovereignty then you cannot own your digital self.

The Trusted Execution Environment is a data sovereignty technology. It's what will allow for the initial phases to take place for an Internet where the users get to not only decide to take back their eyeballs but also to decide which AI they want to feed and for how much. Instead of being prey to AI running in companies set up to mine our data for that AI, we can eat (earn) as we feed the AI.

The flaw with Steem going forward is that it is not prepared for the new ethics of data sovereignty. That being said, Steem is better off than the older tech which didn't reward for the data. The flaw with Steem is that once data is "sold" to the Steem blockchain it is no longer "owned" by the creator. In other words Steem keeps content in the form of "forever records" on the blockchain. For certain purposes this is good, for blogging, for keeping accurate history with time stamps and so on. Any idea for example which someone posts on Steem is pretty much owned by the person first to post it because it's neatly time stamped etc.

The problem is Steem does not allow the data owners to share data in ways where the owner does not lose all control of it. Using Secret Contracts via Enigma it would be possible to share data in a way where no one can copy it and where the AI can use it. I'm deliberately using AI instead of politically correct terms like "machine intelligence" because AI is really just a moving goal post. What was considered AI 20 years ago is now not considered AI anymore, and what we consider AI now is what used to be called AGI.

To be politically correct we could simply say that data is the food of the intelligent machines which eat data to become bigger and more intelligent over time. If you can selectively feed the AI which is aligned with your values then you can benefit more than if companies data mine you, use bots to scrape the Internet or copy your data from all these blockchains. If these AI have to go directly to you to buy bits of you, then you've now created a market not unlike the mutualist biological markets inside the human body.

Tauchain is intended to be what I like to call a decentralized Cyc. It's not specifically designed for "Deep Learning" or the typical forms of AI most people think of but it still requires knowledge to grow. Just as with Steem, there will be a token (AGRS) where if you feed the growth of Tau by providing knowledge to the shared knowledge base then you get rewarded.

Theoretically speaking, we have to decide what is more ethical between the Steem model of put the data in plain text on a public DHT or should the data be separated into "public" and "private" where people can share private data or lease it, for a fee, without having to reveal the data. If data can be monetized without being revealed then I think from an economic perspective it would be an easy win but from an ethical perspective there will be some who want a transparent data commons while others will want data sovereignty.

My current ethical stance is that I want for AI (or companies) to see us as more than "food". I want there to be a mutualism or a mutual benefit relationship. If people are being asked to give a part of themselves to grow some AI do we really want it to be a situation where other people are simply being used by that AI to spy on people to feed the AI? Better to promote trust, promote security, by allowing people to have the choice to participate or not. If AI does not respect self sovereignty in how it captures data (because the companies creating the AI don't) then what could that AI someday evolve into with those kinds of ethics?

In the word "consumer" (which is how a lot of companies see people), a consumer is an eater. If we reverse it upside down then the companies with their AI are the consumers and we are the product being consumed. If those companies are built around feeding the AI then what is the long term ethical trajectory? Maybe we cannot predict that, but we also don't need to bake into our model that our users (who include the majority of us) are to be consumed on behalf of intelligent machines which must eat, and which must grow, and which must consume. The human must eat, the machine must eat (in it's own way), and the machines exist to serve the humans (in my opinion).