Andrew Ng, the founding lead of the Google Brain team, former director of the Stanford Artificial Intelligence Laboratory, and now overall lead of Baidu’s AI team recently told Harvard Business Review what the real potential of AI is and reassured us that it will not take over the world anytime soon.
With years of experience in the industry behind him, Andrew has built many AI products applied by hundreds of millions of people, which has provided him with a solid insight to what AI actually can do. In order to understand the implications for your business, Andrew has cut through the hype and offered Harvard Business Review’s readers an insight to the realistic potential of this technology.
“Surprisingly, despite AI’s breadth of impact, the types of it being deployed are still extremely limited. Almost all of AI’s recent progress is through one type, in which some input data (A) is used to quickly generate some simple response (B)”, Andrew writes.
“Being able to input A and output B will transform many industries. The technical term for building this A→B software is supervised learning. A→B is far from the sentient robots that science fiction has promised us. Human intelligence also does much more than A→B”, he further explains.
These A→B systems that Andrew refers to have experienced rapid improvement, where the best ones are today built on deep learning/deep neural networks that are inspired by the human brain. However, these systems are far from any science fiction, the Stanford Director reassures us. Researchers are frequently exploring other forms of AI where some have proved to be useful in limited contexts with the potential of a breakthrough that could introduce a higher level of intelligence, but this is currently far away.
Andrew mentions that the weakness of today’s supervised learning software is the requirement of large amounts of data needed to make the system work properly. At the time there is a lot of work done by humans that only takes a few seconds, such as examining security videos to detect potential crime or evaluating if a car is about to hit a pedestrian. These activities have the potential of automation, Andrew mentions.
“However, they often fit into a larger context or business process; figuring out these linkages to the rest of your business is also important. AI work requires carefully choosing A and B and providing the necessary data to help the AI figure out the A→B relationship. Choosing A and B creatively has already revolutionized many industries. It is poised to revolutionize many more”, Andrew writes.
He emphasises that after understanding AI’s potential and limits the next step for executives will be to incorporate this into their strategies, understanding its value and where it might be difficult to mimic. Andrew considers the AI community as “remarkably open”, saying how most top researchers now are sharing ideas and open-source codes.
“In this world of open source, the scarce resources are therefore: Data – Among leading AI teams, many can likely replicate others’ software in, at most, 1–2 years. But it is exceedingly difficult to get access to someone else’s data. Thus data, rather than software, is the defensible barrier for many businesses. Talent – Simply downloading and “applying” open-source software to your data won’t work. AI needs to be customised to your business context and data. This is why there is currently a war for the scarce AI talent that can do this work”, Andrew explains.
He provides a balanced picture of the actual potential of AI this far, and reassures the readers that the only fear this far, is AI taking over jobs. Andrew urges all leaders to take this possibility into consideration in the future to prevent job loss that will affect individuals.
This article was first published at: https://hbr.org/2016/11/what-artificial-intelligence-can-and-cant-do-right-now?utm_source=twitter&utm_medium=social&utm_campaign=harvardbiz