Artificial Intelligence is all the rage right now, and it seems companies in every industry are talking about how the magic of AI will change everything. And, when I say every industry, I mean every industry.
Here’s the thing, the potential of AI has been something that data scientists have been touting since the 1970s. This time though, it does feel different. It feels like we are on the verge of AI changing the world forever.
I recently caught up with SugarCRM’s chief product officer Rich Green, a man who has lived through many technology crazes, to get his thoughts on what is different this time around.
Q: What’s Different About AI Movement This Time Around?
Rich: The biggest difference this time: the people who are leading innovation in AI are the same people who are using AI every day. Instead of a university of governmental lab working on AI, you have companies like Facebook and Google who have the deepest pockets, largest data sets and best data scientists. They are the ones working to improve AI and they have the easiest path to integrate what they are working on into today’s world.
There are a couple other big differences. For one, AI related techniques like machine learning get better with more data to interpret. Nowadays, the pool of data is so large that you can expand AI beyond narrow uses cases and do more interesting things, and the statistical accuracy if far greater due to big data. Secondly, the computing power required to do AI has caught up. We have cracked the “AI speed barrier” with hardware and algorithms. AI used to be clumsy and obtrusive, it’s now transparent and can be transparent to people, not having to internalize that many of their connected experiences is powered by AI technology.
Q: I think the industry is still debating what’s really AI and what isn’t. Where do technologies like machine learning, deep learning, neural nets fit into the AI category?
Rich: As you note, AI is a category, not a specific technology. Machine learning, deep learning and neural networks and many other technologies are all part of the AI category. The industry likes to debate what is AI and what isn’t. I argue the sum of technologies fundamentally required to create an evolving intelligent digital assistant, self-learning, self-driving cars and chatbots that incrementally improve their accuracy all fall into the AI category. Tools like Google Translate now use machine learning AI to rapidly improve and provide remarkably accurate translations. In fact in that particular case, the system did something particularly remarkable.
Q: In the past, many people working at the edge of AI technologies have grown disillusioned, and this has stalled progress. Do you think this will happen again?
Rich: That’s highly unlikely. Many of the world’s best AI researchers are no longer confined to pure research in academia. Instead they are spending some or all of their time with Facebook, Google, Amazon and others and are able to leverage the breadth of resources, data and access to accelerate their work. And while that is happening, they can test and validate their work using the largest data and computing engines in the world. With such access and the freedom to experiment and deliver, the pace of innovation is accelerating at an unprecedented pace.
Until recently, there used to be a significant delay in moving research to advanced development and ultimately delivering innovation to a wide range of users. That delay is now compressed because of the tight cycle between research and availability.
Q: Are we currently at peak hype for AI?
Rich: ‘Hype’ is an interesting term. It typically implies that the commentary and the reality are disconnected. Today there is a great deal of discussion and visibility but unlike the past, most of it is either true or will be true quite sooner than most people are able or willing to believe. But we are just scratching the surface of actual capabilities and utility of AI technology. Unlike the 70s and 80s, we are on a very steep slope of growth. There is no logical impediment to this hype cycle.