Why “AI” is more than just a buzzword
The more time I spend engaging with entrepreneurs, investors, and analysts about the startup ecosystem, the more I hear the same old, cliché set of ideas and predictions repeated ad nauseam.
“Tech is a bubble that’s about to burst. Silicon Valley will implode soon.”
“Venture capital is broken. Startups will all turn to crowdfunding soon.”
“[Insert breakthrough innovation here] is just a fad. It’ll be dead soon.”
And so on, and so forth, for years now. One more interesting prediction I’ve heard, though, relates to artificial intelligence. Specifically, a lot of folks are under the impression that AI is just a buzzword, a sticker placed by founders on their companies to give them some measure of differentiation and hype.
It’s certainly a viewpoint worth looking at. Here are the general arguments I’ve heard from people skeptical about the latest wave of AI startup activity:
- Artificial intelligence just sucks. Internet chatbots suck, Amazon Alexa sucks, translation apps suck, facial recognition sucks, etc.
- The term “AI” has become so broad as to be meaningless. Anyone can claim their startup uses AI, or similarly “machine learning”.
- Startups are increasingly using those terms to differentiate their products or services, without actually making advances in those disciplines.
I get the impression that many of these theorists are working backwards, presupposing that AI is doomed to failure and then looking for reasons why. Here’s why I disagree:
Argument: Current applications of AI don’t live up to expectations.
Why it doesn’t work: The root of the problem is in our expectations of what AI-based products should be able to do for us, not the products themselves. “Artificial intelligence” conjures up images of sentience and communicative ability equal to, or exceeding, that of a human mind, images fed to us by science fiction. This vision shouldn’t be the expectation of anybody using an AI-enabled product today, any more than you should expect a Model T to keep up with a Ferrari.
In any case, AI has been improving in leaps and bounds in the past few years, with major players like Amazon, Google, and Tesla continuously developing and releasing new AI-enabled products and services. They may not be perfect, but we have to start somewhere.
Argument: AI and ML have become meaningless as differentiators.
Why it doesn’t work: If there is any reason why artificial intelligence and machine learning have become so common as differentiators for startups, it’s because those tech areas are so versatile that they have disruptive potential in such a wide range of industries. The potential for enhanced productivity via AI-enabled automation alone offers a world of possibilities, to say nothing of the potential applications in home security, transportation, healthcare, cybersecurity, and more:
Argument: Many AI startups aren’t actually advancing those sectors.
Why it doesn’t work: It is important to note that there are two general categories of startups in the AI space: startups using AI to enhance or support their value proposition, and startups developing artificial intelligence systems outright. Though startups of the second variety are more directly innovating in terms of exploring how computer code can resemble what we consider “intelligence,” even startups with only a tangential involvement in AI are still making an impact.
Just as tech companies in the 80s and 90s adopting computers for the first time helped build out the desktop computing ecosystem and drive further innovation, so too are AI-enabled startups proving new applications for the technology, a valuable pursuit in and of itself.
So, what’s the point?
The point is, we should recognize the radical transformative potential of AI, now and in the near future. Here are three arguments for AI as not only a viable area for budding entrepreneurs to jump into, but also an area that incumbents and consumers should watch closely, as they consider how the digital landscape will evolve in the coming years.
AI is one of the biggest areas of startup innovation today. Even as new use cases for AI are being conceived every day, the current set of common AI applications most skeptics point to as signs of AI’s slow growth , namely chatbots using NLP , continue to improve. The goal of AI development — a better approximation for human intelligence — is quite clear compared to those of most other frontier technology areas.
The conditions for better AI continue to improve. AI is theoretically limited only by the amount of power and data applications have access to. Moore’s Law, roughly stating that computing power doubles every two years, has held steady for 50-odd years now. In addition to the explosion of big data tech, this means a very healthy environment for AI development.
Investing in AI is at its all-time high, and is expected to grow. Venture capitalists are investing in more AI startups, and founders are raising larger deals. More funding means more resources with which innovators can develop better AI-enabled solutions.
It’s always good to have some healthy skepticism, but the potential for AI in startups is huge, and as more startups use AI to enhance their products or services, the room for healthy skepticism will need to shrink. Meanwhile, the signs of this sector’s ongoing innovation will persist, and each new life-saving Tesla update, Alexa-enabled consumer appliance, and other new AI device only further validating the need for this sector in 2017 and onward.