Data (Raw or Refined) Wants to Be Proprietary!

You may recall the saying: "Information wants to be free!" But does it, really? This article, strangely enough, is about the inverse of that. Allow me to explain and please bear with me.

A while back, I learned a huge insight ... actually from a candidate in an interview for a lawyer position. What I learned then snowballed into what I believe is a series of broader ideas that were new at least to me. I am sure that others had thought about the same, seemingly decades ago.

THE OLD SCHOOL EU DATABASE DIRECTIVE

I was studying IP law in the mid nineties. It was pretty much the least cool thing you could do at the law school back then. The cool kids were all studying finance law and so on. Given my background, IP and tech law was the only area that I could really get passionate about, uncool as it was. This was the time when the ECJ Magill decision came out and our professor spoke a lot about it. The case involved TV program databases and the conflict between anti-trust and IP laws.

Another incredibly hot topic at the time were the "sui generis" database rights that the EU was working to enact under the EU database directive. If there ever was a "dead end" piece of needless legislation, this turned out to be exactly that. It was supposedly somehow really important, buut not even our esteemed professor could really explain why. There was also this nagging feeling that EU was somehow missing something really important or that the approach was somehow half-hearted. Or both. But I could not pinpoint why that was.

WHAT I LEARNED FROM THE LAWYER CANDIDATE

The earlier mentioned candidate had done the master's thesis on exactly this relic, the EU database directive. A very esoteric subject. Thus, I went about probing a little into that and a lively discussion ensued. In the course of which, the candidate (who must have been a toddler in 1994) taught me that, drum roll, the EU had actually correctly perceived data as the future mega trend already in about 1994 and then proceeded to do something radical: to create an entirely unique, i.e. "sui generis", new type of intellectual property right for databases.

Even more importantly, I learned that this had been done to support and enable the emergence of a new knowledge economy in the EU, exactly based on data and databases. Wow! I had never heard it be put that way. "Europe", "Big Data" and "1994" in the same context...the mind revolts! You know, I was right there at that time, why didn't anyone tell me that this was the idea with all that? I still find that mind-boggling, given that the "data cosmos" today really revolves on the two sides of the I-520 bridge crossing the Lake Washington in Seattle, WA. And certainly not anywhere in the EU. 

What happened here? The EU was truly way ahead of the curve almost 25 years ago. But clearly something went wrong in the execution, or the time and technology simply were not ripe then. After all, the technologies to really mine and monetize data emerged much later in the USA. Or seemingly more precisely partially in the Silicon Valley through the search giant Google, but even more so in my US home town of Seattle, WA, the Cloud Capital of the world, through the cloud giants Amazon and Microsoft.

A VERY HUMAN ERROR

The next realization came to me only later. I can be a little slow in some ways when some (very raw) data is, if you will, processing in the background in the unconscious part of my imperfect brain circuitry. But the idea came.

While the EU had been bold enough to both correctly identify and act on a future mega trend, it had made a very human mistake. Yes, it had over-emphasized the importance of humans for data. Actually, both for creating and processing data. In 2017, now that the era of artificial intelligence and machine learning is finally here, it is easy to be smart in hindsight. While I can recall the nagging feeling that something was wrong with the EU database directive, I could not pin it down at the time. But it indeed seems so very obvious now.

While not as human centric as copyright law, which centers on human authors and their creativity and "sweat of the brow", the database rights do not require quite such creativity, but they still require "substantial investment in either the obtaining, verification or presentation" of the data. Different, yes, but still centered on the human effort, even if perhaps partially assisted by technology.

Even more importantly, the database rights clearly do not apply to what in many ways is the most valuable form of data, namely raw data. This in fact is the exact opposite of the perfectly organized databases that the EU databse directive contemplated as being important. (Raw) data can be mined and refined in different ways, some of which might not be even technically feasible at the time the (raw) data is created – but only much later perhaps. The value arising out of raw data is, a little bit at least, like the value of derivatives in high finance. In that you can derive all kinds of valuable and useful insights and sometimes even patentable innovations from tons of simple raw data. Only that the AI does it for you. Fast.

So, a very human error was made, i.e. the EU overemphasized the importance of humans for both collecting and refining the data. And, therefore, also ignored the value of raw data entirely.

THE VALUE OF (RAW) DATA

Next I came to think of all the different contexts, where I have heard that (raw) data is super valuable and mission critical over the recent years. At Microsoft in Redmond, you surely were exposed to the notion of the data in general having value, even if perhaps extractable only later. Some of you have surely overheard even statements like "Data is the new oil." And without going deep into the weeds here, it obviously is.

If you get (raw) data from, say, a car (really through connected sensors in the car) this then enables very many worthwhile commercial activities, like: (i) improving cars, engines, safety systems; (ii) selling maintenance services and spare parts (which is really what the car makers make their money from these days, as e.g. any Mercedes customer will appreciate); (iii) selling media content to cars and people in them; (iv) selling really anything from coffee to parking to ski tickets; and (iv) countless other business activities that will emerge over time.

So, yes, (raw) data is a little similar to (raw) oil such that you can make tons of different things out of it. Just as (raw) oil is used for anything from cosmetics to medicines to plastics to gasoline. Ironically, latest today I spent an entire hour discussing value and treatment of "big data" in the context of a complex IP centric M&A project.

Something is changing with data.

A little like computer software in the early 1980s. You can sense the same exact pattern here. An era of change may be upon us, also legally speaking.

(RAW) DATA WANTS TO BE PROPRIETARY

So, if (raw) data is so incredibly valuable, how do you own it? After all, ownership is required to really enable things like securitization, using as collateral, selling, buying, inheriting, licensing etc. In other words, all the things that make the modern economy run. And, in fact, even for making things free in a principled manner (e.g., open source software, which is largely based on copyright).

If you doubt my words, please familiarize yourself with the great Latin American economist Hernando de Soto (for those who care, he and not Nalle Wahlroos was also the favorite economist of the late "TL").

Property undefined and unprotected is property neglected and destroyed.

https://en.wikipedia.org/wiki/Hernando_de_Soto_Polar .

The relic in the form of EU sui generis database rights clearly is not the answer.

Personal data protection is not the answer either, as only a small subset of (raw) data with value has anything to it that can necessarily be connected with individuals. Actually, most of the (raw) data that has value is generated by machines (think IoT) without active human involvement and then later also processed without human involvement. Why all this talk about privacy everywhere then? Well, lawyers like to talk about the laws we have, not about the ones we should have. An easier topic.

So, while I applaud the efforts of the EU and others to protect right to privacy and personal data as a human right, of sorts at least, it really is not the solution either. Also, privacy is one of those virtues, which while important for the individual, is not really a monetary value. It is quite like freedom instead. But that will not be the solution for bringing the (raw) data into the sphere of commerce in a controlled manner.

As an example, raw ASCII form data indicating mere temperature precisely through a dense network of sensors has commercial value and can be refined into a ton of things, but there really is nothing personal about simple temperature data sent by thermometers. Quite the opposite in fact. The weather is what it is, irrespective of us humans. You can try a rain dance, if you disagree. There are many other examples.

Patents are obviously not the answer either, as they only involve novel and inventive inventions that at least should be the exclusive domain of the nobel laureate, the professor, the PhD or at least some other "person skilled in the art".

Copyright equally cannot be the answer, again as it overemphasizes the human creative effort and the "sweat of the brow". Not to even mention trademarks that really protect, often irrational, emotional associations between services or products and their commercial sources.

Contracts can offer some instruments for creating schemes similar to property rights, but are inherently "unstable", private (not public) and complex.

What about trade secrets you might ask then? Well, not really. While there are different topical developments in that area, much of the (raw) data is not only created by machines to be refined by machines for insights, but also too crude and simple to constitute trade secrets. Though there are exceptions to that.

The value with (raw) data largely vests in aggregation, sheer scale and observable patterns, but the underlying (raw) data can actually be really simple. While still hugely valuable.

THE NEED FOR A NEW TYPE OF PROPERTY RIGHT?

So, query the following: Do we need an entirely new type of property right for (raw) data, even if simple, to truly allow for it to be transacted, securitized, collateralized, inherited, owned, pledged etc. like other types of property? And even to be made free in a principled manner like open source software.

The great Hernando de Soto would likely say: yes. 

If so, how should we account for the fact that much of the data is created by machines for machines? With very little human involvement. Who would own that data? The human owner of a physcial IoT device or perhaps even the "owner" only of some static IP address (the Internet Protocol in this context, not intellectual property)? The crucial point being that, if you want to encourage investment and activity, there needs to be solid protection for the rights resulting from that investment and activity.

Of course, here that paradox is that the activity to be encouraged entails humans developing and investing in machines that both collect the data and mine it for value. But it really is the machines and AI doing the heavy lifting in terms of creative and inventive activity at scale. In fact, you could also say that, in overemphasizing the value of organized data(bases), the EU failed to appreciate the power of raw computing. These days, you can organize and analyze vast pools of data in microseconds. And do it all over in different ways. So, computing power is becoming a commodity, while the data often becomes the bottleneck.

And, as always when the law lags behind societal progress, the engineers hack a solution. Currently, it centers around owning and controlling the media, on which the data is stored. As well as access to such media, of course. But that is a rather crude approach, not too unlike software being protected through owning physical media (and as trade secrets also) before copyright protection was extended also to software. This is why I think that there is still some under investment in "big data". 

"Big data", for the time bing, remains a little like "sex in high school" such that, while everyone talks about it, very few actually do it.

Anyhow, it was fascinating to realize, based on a lawyer interview, that the EU was on the right track already in 1994, but made the very human error of overemphasizing the importance of us, the humans.

Though, in their defense, all forms of private property still roll up to humans, be it in the role of a landowner, shareholder in a company, copyright holder, owner of a coin or banknote, owner of a gold bar etc. And maybe that is the biggest paradigm shift here…maybe this is the first time in the history when a new class of valuable property will be created and refined purely be machines (which still, for now, are owned by humans or at least human shareholders in companies owning the machines).

Put differently, I still have this creeping suspicion that I am missing something important. It just might be that while data wants to be proprietary that may not mean being proprietary, directly or indirectly, to a human.