A sympathetic friend can be quite as dear as a brother — Homer
We live in an interesting world. Nobody knows what happens next. Even if they think they do, they don’t know how soon. Take the case of Artificial Intelligence (AI). AI has always progressed in spurts. There were two AI winters where research and investment in AI waned. In the past decade, we have seen a resurgence and increased interest in AI. Today, most of the applications of AI are confined to solving lower order problems compared to what AI is potentially capable of. However, the world is swinging between observation and paranoia as it concocts various scenarios about how AI will affect the world. This article offers three frameworks to look at a world where AI will be accessible by both the large Goliaths (Facebook, Google and the ilk) as well as by common men as it gets rapidly democratized.
There is a heated debate on many issues surrounding AI i.e. what constitutes AI?, how do we use AI?, how do we ensure data privacy without stifling innovation?, what are economic, moral and ethical implications of AI? Right now, it seems like many blind men are touching different parts of the elephant and defining the problem from their viewpoint. Harmonization of approaches will require a common mission. In human terms, that mission is an existential crisis. We are not there yet. Therefore, the debate rages on and holds the world in its sway. There a few different ways of looking at AI outlined below.
Reasoning By Analogy
If data is the new fuel, then platforms such as Facebook, Amazon, Google, Baidu, Tencent or Alibaba are the next OPEC. The difference between the OPEC (Oil Producing and Exporting Countries) which together control 82% of the worlds oil reserves and the big 9 AI companies (Baidu, Tencent, Alibaba, Facebook, Google, Apple, IBM, Amazon and Microsoft) is that the data that is residing within and feeding the algorithms of the big 9 companies will come not just from their home countries but globally.
This is because the best way to remove bias out of algorithms is to increase the sample size and create a very diverse sample size.In that sense, these platforms have crossed nation state boundaries in search of profits. This behavior brings them in the cross hairs of regulators (guardians) of the world. Obviously, the interests of local citizens residing within these nation states clashes with the profit motives of these platforms.
Ideally, anybody using personal data, including the AI Goliaths, have to explain how they are using the data. I use the word anybody because exponential technologies are getting democratized. Therefore, if we focus solely on the big players, we lose focus on rogue players that may act alone because they have access to exponential technologies such as CRISPR.
Countries have taken different approaches. The European Union has decided to regulate data privacy. US has a less regulated approach. While India has thrown open the data on its citizens anonymously in the public domain, China has a less open approach. It is also looking at facial recognition technologies to track citizens. Thus, nations across the world have different approaches to data. These approaches are not likely to be harmonized unless the threat is perceived to be existential.
At The Crossroads of Exponential Technologies
One of the most beneficial applications of AI is its application towards improving human health, longevity and ability to deal with chronic conditions. If we assume anonymous medical data from all across the world can be pooled together- which is big if-we can treat chronic conditions or prevent chronic conditions from arising. On the other end of the spectrum is the ability to play god i.e. use CRISPR, RNA interference or other gene editing techniques to simply get rid of the gene responsible for the chronic condition when the baby is born. Of course, nature has a way of getting back at us. If we do play god, it can have catastrophic unintended side effects. On balance then, using AI to understand the inner workings of the human body seems like a better approach than gene editing.
Another example of cross roads in exponential technologies is the quest for more accurate data sets which fuel AI algorithms and cybersecurity. How do we ensure anonymity of data if the data is susceptible to hacking at the source. There are many technologies that are advancing simultaneously. These technologies include AI, Quantum Computing, Genetic Editing, Blockchains and
Leaders Have To Be The Guardians
There are certain truisms that can guide the way we think through our approach to Artificial Intelligence (AI) and to really any exponential technologies affecting our lives today and in the future:
If there is one thing we know for certain, it is the fact that everyone is driven by incentives. The private sector by profit, the government (hopefully) by the welfare of its people and rogue actors by pure self-interest. The problem with exponential technologies today is that they are getting democratized faster. The dilemma that we face, therefore, is that how many people’s behavior can we control. The answer certainly is not everyone’s. A look at the history of hacking will tell you how a single individual can change the destiny of corporations. Another case in point is the extremely dramatic announcement that a Chinese researcher — He Jiankui claimed to have created a genetically edited baby. By extension, we can expect that there will be rogue actors as far as any technology is concerned.
It is almost impossible to predict how fast AI will progress. Right now, AI is using limited data sets and is directed towards specific applications such as process automation, facial and voice recognition, autonomous vehicles. All of these applications optimize for a specific objective function. Therefore, an argument can be made that we should not be scared. Yet. However, many futurists seem to think otherwise. There is a wide chasm of difference in opinions which is difficult to bridge because nobody is aware of all of the AI efforts worldwide.
Regulation, traditionally has to catch up to innovation in the private sector. If regulation precedes innovation, it will stifle entrepreneurship. So , you need to give space to innovation while avoiding the pitfalls.
All of these truisms put together have one inescapable conclusion. As Peter Parkers’ uncle in Spiderman famously said ‘with great power comes great responsibility’, every leader in the AI world has to be the guardian. The leaders are the first line of defense. Government regulation has to act as a partner sometimes and as the second line of defense other times. Every leader has to treat access to data as a fiduciary responsibility. No data is worth selling out our humanity for. Not every individual will understand that. That is just the nature of the human race. Therefore, if things go rogue, we will have to band together as a specie to combat the threat. For now, in the interest of innovation, we have to let the evolution of AI play out with limited restraints and disparate approaches. The plan B should be a faith in our capacity to come together. This faith is by no means misguided as history will tell you.