Why Economist is wrong about AI

A recent Economist article argues that the big five companies (Alphabet, Amazon, Apple, Facebook, Microsoft) have access to dispropotionate amounts of data and their data should be regulated for access to more companies. This is because with access to so much data, only these five companies will dominate the AI field and AI applications (access to large amounts of data is required for effective AI) .

All three arguments made above are completely flawed and I’ll counter them. The author’s flawed arguments are:

1. These are the only companies with large amounts of data
2. These are the only companies which will dominate AI
3. Government regulations are required to share data so others can also build effective AI solution

First, these are not the only companies with access to huge swarths of data. Most cloud based enterprise solution providers have amassed large quantities of valuable data. Companies like SAP, Oracle, Salesforce, Workday, etc. have more data about manufacturing processes, corporate financials, HR, etc. than anyone else. Industry vertical solution providers specialize in data for their respective industries. Example — Veeva has clinical data, Guidewire has insurance data, and Ellie Mae has home mortgage data. In fact, leading solution vendors for a given industry has a lot of industry relevant data and can easily augment the missing data by combining it with data from “data brokers”.

Data brokers like Data Republic, Nielson, Experian, Dawex, have created a Data Marketplace for anyone to buy or sell data. This makes all kinds of data accessible.

Real issue is that most companies have not yet figured out how to leverage their data to build effective AI solutions. The issue is not, and will never be data availability but rather the ability to build useful AI applications.

Second, These companies will not be the only ones to dominate AI. In consumer space, perhaps. But that is a myopic view into the world. Businesses will spend much more on intelligent applications as vendors learn to incorporate intelligence (AI) into their applications. This will produce a wide variety of AI based solutions. In the same article, author points to $7.6B investment in 2017 alone in AI. These investments are obviously not in the big five mentioned above. Actually, here is the PitchBook link to investment areas. So clearly there is a an existing thriving ecosystem of AI based solution providers and there is a potential for several more.

Third, government regulation is the last thing that will enable innovation. When the author sites the european example:

“Europe’s impending data-protection rules, for example, require firms to get explicit consent for how they use data and to make it easier for customers to transfer their information to other providers”

we are really talking about allowing individual users to port their data, not forcing data sharing en masse. Moreover, I am not even sure if Europe strikes as a good example for AI innovation hotbed.

Jerry Chen of Grey Lock published a complelling post about Data as a competitive moat. I subscribe to this strategy. Most organizations in the capitalist world will not share their competitive advantage or let it be regulated. Regulation of competitive advantage is a naive stance and an Economist article professing this, is disheartening at best.

To summarize, I believe there are ample opportunities for innovation in AI across a broad spectrum. There is no shortage of data or business potential. Rather, companies have to be willing to make AI a part of their strategy and invest in it now! Executives have to educate themselves on the potential of AI. As with all innovation — planning, foresight and patience is required and I can only think of one quote.

I will prepare and some day my chance will come

— Abraham Lincoln