Over the last five years we have come to be able to track anything, and the costs of doing so are shrinking. Just as Moore’s law predicted the doubling of computational power every 18 months (and therefore the cost-per-computation halving every 18 months), we are seeing the emergence of a new era of cost reduction: cost-per-prediction.
This is thanks to two parallel waves in technology: Artificial intelligence (AI) and the Internet of Things (IoT). AI is producing loads of new information due to its ability to turn unmanageable data into manageable data (see Pt. 1), or noise into signal; and the IoT is putting that capability everywhere via connected sensors embedded throughout the physical world and turning things “smart”.
We have smart transport, smart grids and smart buildings conjoining into smart cities. In these spaces, everything is being sensed and understood. We usually monitor traffic using cables laid across a street that mark speed every time they are run over by a car. With AI, we can instead use a street camera and additionally track: brand, make, model, speed, pedestrians, weather, etc.
When the formerly un-trackable becomes trackable
Most of our social contracts are written based on the assumption that public spaces are full of noise, and that our lives as pedestrians blend in with that noise in relative privacy. We further believe our lives at home are fully private, outside the awareness of others and not even a part of the noise.
But, is that still a reasonable expectation when we own connected appliances that are always on? Amazon’s Alexa, Apple’s Siri, Microsoft’s Cortana, the TV and the thermostat all listen to 100% of what we say—presumably for a command, but nonetheless listening. No need to even consider satellites watching everything under the sun or cell towers tracking our location when these all-hearing technologies occupy our most private spaces.
Today, AI is processing all that collected data into information that could be useful in all sorts of other domains. NVIDIA has rolled out its Metropolis platform for AI-powered video analysis that connects cameras on each street corner to know and understand everything they see. With that information, governments and businesses are able to engineer lives and societies even more comfortable than they already are, comfortable enough that we’ll keep giving away our information for access.
However, the benefits of smart homes or cities don’t change the accompanying vulnerabilities. Security is an obvious concern. Consider the DDoS attacks carried out using connected printers and baby monitors to crash internet servers covering the eastern US and Canada. Those attacks exploited the poor security of those devices, and some manufacturers have started fixing those problems. Even so, what about those who have legitimate access to the information those sensors produce, and whose interests conflict with mine?
Valuing data means valuing privacy
Currently, there is no accounting of what data-generating companies know about us. Yet, The Economist is saying it’s more valuable than oil. This, just like the financials of a public company or the value of intellectual property, should be transparent for stakeholders (including those who are being tracked) and governments through disclosures and auditing. Yet data is nowhere on the balance sheet; so, how do the data aggregators like Facebook or Google even assign ownership or value to that data?
Oil companies are taxed all along their value chain, including where the oil is originally extracted. Sure, they invest heavily in infrastructure, but governments still expect a large amount of that generated wealth to be redistributed. We as a society tend to look down upon companies that extract resources from countries without offering a return beyond the infrastructure they create, while monetizing and paying tax elsewhere. Well, we’re seeing just that with AI's conversion of sensor data into information: the creation of new intellectual property that flows across borders to be monetized elsewhere, escaping tax until it’s further down the value chain.
We need pro-active legislation that enforces data governance for this wave of technology. Just like any other asset, data (public and private) ought to be inventoried and valued on the balance sheet. The value of that data should then be taxed like any other asset.
Accounting for data's value will not only lead to critical tax revenue, but it may be the best way to also enforce regulations on sensitive information and ensure proper data governance.
A new social contract
Unregulated, AI will only benefit the few. The value and sensitivity of data will require a new social contract for us to move smoothly forward into this new economy of information.
FDR increased the role of the state in protecting U.S. citizens when he created financial safety nets through social security after the great depression. Today we are hearing similar ideas like universal basic income as a means of raising the floor before the bottom falls out entirely from automation. I believe we should also double down on providing continuing education to correct for those spikes and help people re-enter the workforce.
Funding these initiatives is not easy. Politicians can be hard to convince to spend more on education. I think it’s because we aren’t capitalizing the overall social costs; we’re too caught up in the short-term and missing the lifetime value. That lifetime value without education will take an even bigger hit when automation increases. When will it be enough of a benefit to invest in education—or a cost not to?
We can’t wait and see
Typically governments have been reactive, waiting for situations to arise and shift public opinion. But now, they can’t afford to wait. They need to be proactive.
There is a dramatically different government in the coming future, politics are ripe for upheaval. People are coming to realize they don’t have control of their data, and are going to look to put their vote behind someone who can deliver security, without taking away the benefits.
Not only that, there’s a serious financial incentive. The potential tax revenue that comes with regulating data is massive, and much more effective than taxing the robots as Bill Gates has suggested. It also makes a new social contract and healthy innovation viable.