The AI Hype
03.07.2017

As it appeared in huffingtonpost.com, by Pascal Kaufmann
July 03, 2017

2016 saw a record number of AI companies be born and 2017 shows no signs of slowing down. Recently Alphabet, the parent company of Google, announced it will be repositioning itself as “an AI company,” and Apple Inc. just unveiled plans for its Neural Engine chip for iPhone AI. It could be said, that we have moved from the Digital Age into the Age of AI. Perhaps there is still one gleaming error involved in this conclusion here– Do we actually know what AI is?

What everyone is labeling AI (what the investors are going crazy for)—actually has little to do with what most people would define as “intelligent.” Arguable, AI as it stands currently is fast and powerful computers which are incredibly good at processing data. To call them intelligent though, may be us getting a little ahead of where we actually are in discovering what Artificial Intelligence truly is. Take for example, calling a watch intelligent because of the tinned intelligence that a mechanical engineer has put into the device. Most brain researchers would agree, that today’s “AI” is not much more than tinned human thought compressed into source code and impressive artefacts and devices.

Weak AI vs. Strong AI

The distance between what people call Weak AI and Strong AI is boundless. Weak AI is when a computer, after scanning thousands of pictures of cats is able to identify a single picture of a cat. Strong AI is when a computer is able to identify a cat by glancing at just a tail. Weak AI depends on rules for a specific task, Strong AI processes data beyond rules and mimics human intelligence.

The real problem with Weak AI, is that it is mainly statistics. It may develop better rule sets and process data faster, but it is limited by the very rules it is based upon. People, actual humans, don’t think that way. If we are basing the development of AI on the expectations of human intelligence, then we first must understand how human intelligence works, and to do that we must understand how the brain works. Powerful computers will not achieve the investment, business, and societal expectations currently associated with Weak AI.

We Don’t Know How the Brain Works…Yet

Systematic scientific studies on how the brain works and tackling the brain code has only really been in place for the past 50 years or so. From those 50 years, we know surprisingly little. We have a very basic understanding of how neurons work. While neurons only make up for about 10% of the brain. 90% of the brain consists of other cells whose function are unclear at best. With so little known about the brain, it may seem like a more worthwhile endeavor to focus on Weak AI and just reap the limited benefits of that. Putting aside the scientific challenge, there are a lot of business applications when it comes to automatization and industrialization 4.0, such as real automated virtual assistants, virtual lawyers, bankers and so on.

The difference from 50 years ago is that we now have the technology to not only look at the brain better but to unite the scientific community better and faster. These advancements of a greater clinical ability to look at the brain and a more efficient approach to boundless global collaboration, means we may be capable of cracking this intelligence puzzle within our lifetime.

Embracing the Latest Technology as A Species

If we really want to have machines able to surpass human intelligence, the investment and time needs to be put into cracking the brain code, meaning the underlying principles of neural tissue. This does not mean, that we need to understand every individual cell within the brain. Much like the movements of planets and stars became clear after Newton described the underlying principles in 1687, a defined set of rules may be sufficient to unlock the dynamics of the brain, referred to as “Brain code”.

True AI which mimics the brain will bring about everything from personal cyborg upgrades and sentient beings to virtual human. AI will be more than beating a master in GO or a poker game, it will be AI which is able to master, enhance, and even invent a multitude of skills and solutions. We will be able to develop computers which enhance our own intelligence and abilities. The key to all of the potential which startups and investors see in this tech is not through statistical brute force approaches but through cracking the brain code and applying that knowledge to how we develop Strong AI.