Privacy is a public concern for over a hundred years. But it seems each year we lose more of our privacy. Today thousands of different companies hold data of most of us. Smaller companies have limited data-tools and therefore invade our privacy in a small way. But Big Tech can recombine thousands of small bits of data from us to detailed profiles and use them to seduce us to choices we would not have made on a voluntary basis.
Apart from privacy, two other features make data very important:
1. Data present a more significant part of the value in the modern economy.
2. Data are an essential input for the Artificial Intelligence systems that will shape our future.
Today, and even more so in the future, the power is with the parties that hold most and best data. Those data make a few companies and nations very rich at the expense of other countries and common citizens.
For the above reason, we believe that data should not be in the hands of the institution that collects the data (for example social media platform). In principle, the data should belong to the ones producing the data (for example the user uploading content to the media platform) or the ones to whom the data refer to.
There are some disadvantages by declaring that the original producers/subjects should own data:
1. When a raw piece of information comes into a system, it is difficult to decide about the rightful owner without proper investigation. Doing this real-time correct for large streams of data will be almost impossible. There is no simple way to quickly decide the rightful owner(s) of each piece of information.
2. Many data have multiple subjects. If the data is a video clip of children in the park, the stakeholders of this video are the person taking the clip, the children on the clip and the parents of the children as legal guardians.
3. Sometimes society might decide that certain pieces of information should not be controlled by the subject producing the data. Some examples are data from terrorists or paedophiles.
The European GDPR directive puts the ownership of data already in the hands of the original owner. In practice, users cannot exercise their rights. As an individual, you only have the choice to accept the terms and conditions or decline. But if the data are owned instead by a user-organisation, the management of this organisation can negotiate a proper reward and an acceptable level of privacy with large collectors of data.
The golden bullet solution
This is a relatively easy and very effective solution: Give the legal ownership of data to a data-cooperative. The members own the cooperative collectively. The cooperative can give the usage of data to the rightful user. It may also decide to delete specific data on the request of stakeholders. By putting the decisions on what to do with data in the hands of the cooperative, there is an independent party taking a decision.
Collectors of data will never give up their valuable data voluntarily. However, a government can decide that data-cooperatives must own all data collected from its citizens or in its territory.
There is a performance issue. Moving all data from collectors to cooperatives would be hugely expensive for both the collector and data-cooperative. The computer-systems of collectors would become very slow. Imagine you are checking your Facebook timeline and Facebook would have to download each item from a distant server of the cooperative. It would genuinely kill the user experience.
There is a cost-effective solution: Build an API between the collector and the data cooperation. (An API connects two or more software systems.) It will not affect the speed of computers at the collectors and might increase the cost-price of collectors by a small percentage, but not much.
Through the API, a data-cooperative can access the data of the collector at any time.
A data-cooperative is much better at taking care of the privacy of users. The collector has to comply with the local regulation, but its financial interest is in evading this where possible to increase its profitability.
The users govern the data-cooperative so that it will put the interest of its users above the profit for any party. The data-cooperative can make an easy to understand back-office for its users where users can see in detail which data it has stored from them.
The data-cooperative can install a member-approved model, with binding instructions for producers, collectors and third parties for the use of data. It can make detailed guidelines on data-level and meta-data level and show producer of data (consumers) how its data are used and secured in an easy-to-use backend environment. The control of data is back with the citizens. Those measures can bring back the privacy-level of ordinary people to a level not seen any more in the past decades.
Both in the case of the current situation where data are with collectors and the new setup with data-cooperatives, governments will probably want to be able to check on data of its citizens up to a certain extent. The government has this privilege. In essence, nothing will change at this point, except that citizens might be more aware of the reach of the government in case of a data-cooperative.
Sale of data
In case the cooperative sells data, the income is for the cooperative. It will not be dependent on gifts of government support. It can re-distribute the profit after expenses to its members. In this way, the people collect the benefit of the data. Now the collectors (Big Tech) take most of the gain.
The current model where the collectors are owners of the data has one advantage: The collectors can offer lots of fantastic services for free. Everybody is now used to free search-engines, free social media, free email, free navigation, etc. The downside is that SOMEBODY MUST FOOT THE BILL! All those beautiful services are now (with often substantial profit margins) paid by selling data. The most elegant solution would be that the data-cooperative negotiates with the collector that he can re-use part of the data for free, in return for free services to its customers. Apart from that, we believe every collector should offer a subscription option for the users that want full privacy and therefore nothing can be earned from their data.
Education is needed for the public to make them understand that they are selling their privacy in return for free services. Most consumers do not understand this trade-off. The task for this education lies both with the government and collectors of data.
The decision what to do with data is in principle in the hands of the data-cooperative. But the provider of the data should be allowed discuss with the cooperative how to handle the data. A data-collector should have the possibility to go to court in case it feels the data-cooperative is not treating it or its data appropriately.
In a traditional cooperative, members that contribute more to the cooperative, receive more money. Likewise, producers of valuable data should receive a proportionally part of the profit. Producers that want to keep most data private will generate less revenue and should get proportionally less paid.
Usually, data will have considerable value, but there might be data-collectors where data have no or little use. In this case, the legal ownership will in principle still be with the data-cooperative, but the API will be only used to give access to the producers. The law should put the cost of the API on the collector of data, so the data-cooperative is not financially harmed.
Implementation of data-cooperatives
Legislation of data-cooperatives is of a national character. But large collectors of data will be global companies. So it would be practical if data-cooperatives would cover multiple countries, if possible with a synchronised legal framework.
If for each data-collector there would be one data-cooperative, this would result in an inefficient and costly arrangement. One data-cooperative for multiple collectors would hugely reduce costs. Forcing collectors to give access to numerous data-cooperatives is counterproductive and the chances for security breaches.
When countries start with data-cooperatives, it would be wise to make this only compulsory for huge companies (for example a threshold of 10-billion-dollar turnover worldwide). Smaller companies can be added, after a successful implementation of the first data-cooperatives.
At the start of the internet age, there was a flurry of new promising start-ups. Now the dust is settling, and large companies take the lead. The data-cooperatives might unleash a new wave of innovation and start-ups when data, the digital currency of the modern age, are available for new aspiring entrepreneurs. The data-cooperative will convert the oligopolies of today into a free market, where the invisible hand of Adam Smith will result in many suppliers each earning a reasonable profit. Instead of a few billionaires, there will be many millionaires.
In 2018 several of the largest organisations in the world are the leading developers of AI systems, and they can and do use the data that their users have legally signed away. In a system with data cooperatives, they will have to ask permission and pay the price to user-controlled data cooperatives for feeding data into AI systems. The members of the cooperative have the right to decide to allow its data to be used to feed into AI systems or not and on what conditions, so we together will get the chance to choose about the character of future AI systems.