# Privacy, data ownership and AI By [f](https://paragraph.com/@maybearmed) · 2024-01-30 --- One of the most important realizations in the future will be that the privacy narrative was never fundamentally about real world protection from physical entities but instead that it will primarily serve the use case of giving us the ability to own our data just as we would own any other asset, and that this will be made clear in the face of ever escalating usage of AI agents (we will concern ourselves with non-autonomous agents which are under human control, we will call them simply “agents”). In order to explain this more clearly, we will start by quickly defining what use resources are to us as humans and where data fits into this picture.The ultimate motivation for humans is self-preservation. For this reason, we must acquire resources which allow us to sustain ourselves. All resources are physical, but some of them are abstract in the sense that we must ascribe additional meaning to them in order to make use of them. Data is such a resource. We cannot eat data, but we can use data to get food, money, and all the rest of it if we know how to properly interpret and make use of it. For this reason, some of us then create agents to help us acquire these resources, and those agents require **data** in order to be sufficiently competitive as to actually show capability of generating net positive returns. Agents also require physical resources such as **computing power**, **hardware**, which is used for training and workflow execution, but also further data such as the actual **program instructions (the agent’s code)** which define the agent. In any case, the most important realization is that out of all of the resources formerly mentioned, which are required to sustain a functional agent, **data is the only** **resource** **for which** not only are proper fundamental rights of ownership not respected, but also, **the idea of ownership has fundamentally not entered the mainstream**. In general, humans understand ownership over tangible physical goods and abstract value representations, yet still, the average human is not informed about the concept of data ownership which we have laid out. #### Why have humans not fully adopted data ownership? Very similar to any historical developments, as long as humans are not forced to modify their behaviour due to changes in the abstract economic and social systems they are interacting with, they are not going to invest extra effort into predicting future developments and acting on them in forward. In general, for both beneficial and unfavourable events, when we consider both the future and the past, the time until such an event will occur, or the time since this event has occurred, must be relatively short in order to cause humans to modify their behaviour in advance, or modify their behaviour as effect of. These are respectively scaled by the expected or the realized intensity of an event, and as such this also has to be considered when trying to arrive at a conclusion when a narrative could develop. #### Do humans currently appropriately judge the possible severity of not modifying their behaviour such that it includes and acts on the concept of data ownership in the face of a surge of autonomous agents in various competitive environments, notably markets? We have formerly made the assumption that most humans are _essentially_ not working on enacting and defending the ownership over their data, in other words we could say that data is the only the resource over which most humans have not yet learned to properly enact ownership over, the resource which at the same time is key for agents which are set to overtake many competitive environments. We are going to focus on markets, and note that agents are able to process data many orders of magnitude faster than any individual. Further, agents then make use of this data to extract value from the same counterparties they have been harvesting data from. Since all participants in economic markets currently interact with these markets electronically, and ignoring all would-be temporary bounds, most market participants are set to lose all possible advantages in markets exactly for the reason that they share those sets of data, which should for _non-autonomous_ agents be much more difficult to acquire, for free.At the same time, it is to note that agents are by many orders of magnitude faster at data processing, and that any possible personal or collective effort of improving market participant coordination and data processing capabilities is not going to be sufficient enough to stay competitive. Data harvesting left unchecked, due to the superior execution ability of agents, will logically lead to endless value extraction from most market participants. #### How does data ownership help? It becomes clear that the only possible way most market participants will be able to preserve their competitiveness will not be in enhancing their data processing capabilities, but rather in protecting the data which enables them to execute actions. Since we are talking in a non-autonomous scenario, we are assuming that many of the knobs when it comes to market liquidity, project legitimization and further, ultimately lie in the hands of humans. It is thus not doubtful that the general interest will be in preserving ownership over valuable data which gives information in advance about changes, or volatility, in areas of interest. It is to be expected that since almost all forms of digital identity schemes are as so far incomplete, that either only personal or contractual, meaning legally obligated with NDA's and all the other terminology, types of information sharing will be the ones to survive. All other forms of sharing information from which participants may confer advantages, as we commonly see today, are ultimately doomed since they will cease once it becomes obvious that these negatively impact all market participants because of the involvement of agents. The effective data utilization of an agent is by many orders of magnitude, in tandem with the data processing capabilities, larger than for a human. Thus, the involvement of agents is alike to a dark forest scenario, in which any data publicly published immediately undergoes analysis after which execution based on this data follows. There is no longer a time delay if information is shared publicly, for this reason it must be obfuscated, meaning participants must be able to enact ownership over their data otherwise they will not even be able to properly utilize it. #### What are the means to data ownership? The means to data ownership lie in technologies which transform data into those other forms of data, which allow the owner to selectively allow or prohibit parties from engaging with it and which carry the same informational content as the originating data, even if it has not been recorded on the same medium. Thus, cryptographic technologies are the primary instrument of enacting data ownership, the use of which will increase in tandem with the increasing number of deployed agents and their capabilities. We know that technologies used for the obfuscation of data using cryptography are also called _privacy enhancing technologies_. Thus, any discussions about privacy are actually discussions about data ownership.Going further, the exact consequence of the application of cryptography is the formalization of the criteria for access to data, and thus it also enables more complex pricing and trading of data. We have already seen forms of criteria formalization and pricing of data in a number of projects, over time the markets will move even more liquidity towards such projects as they start entering the mainstream. Even though these are _technologically_ bad examples, projects such as [friend.tech](https://www.friend.tech/), which essentially offer content creators the ability to sell access to personal chats, represent exactly this concept of access to data with automatic pricing _based_ on access. Still, the data ownership of the content creator in terms of what the platform has access to is obviously fully violated and instead cryptographic primitives should preferably be used for sharing access. #### Conclusion Understanding the importance of data ownership and knowing how to utilize the instruments of data ownership is going to be key in navigating markets and preserving your future competitiveness, for this reason you should familiarize yourself with “privacy” (read: data ownership) enhancing technologies, and readily make use of them. --- *Originally published on [f](https://paragraph.com/@maybearmed/privacy-data-ownership-and-ai)*