A Conversation With Our 2018 Keynote Speaker: Michael Rogers, MSNBC’s “The Practical Futurist” (Part 2)
A lot of people worry that bots will “steal” jobs from human workers. What are your thoughts on this issue?
My thoughts are somewhat mixed here, because I think this is a very real concern. I’ve worked with lots of companies, and we're already seeing AI and automation begin to eat away at some of the entry level jobs in white-collar professions such as law, accounting, advertising. These were fields where you would bring in the youngest workers and give them routine kinds of work. In law, for example, associates might work on research, discovery, or contract drafting for a few years before they were ready to go out and meet clients. Today, a lot of those jobs can be done significantly more cost effectively and efficiently by AI and cognitive computing. Now law firms are beginning to rethink how to keep their young associates busy while they're still learning the ropes. It's not so much that AI steals jobs. It simply does them better and, as a result, is changing the nature of work. But this fear of a net loss in future jobs is not something to be discounted. The notion that when technology takes away jobs, it creates new jobs, has always been true in the past. However, it's never been tested in this situation where we are taking away cognitive jobs, the highest functions that humans can perform. I think everyone involved with intelligent technology needs to think about this issue. Those new jobs will not appear magically on their own. We really need to think about how AI will impact our workforce. And the other piece is how do we train workers now? What skillsets do they need as they move into an environment where they are literally collaborating with AI all day long? What skills should we be teaching our kids- what skills will a computer never be able to master? That's one of the big questions.
Why is AI bias an important issue and what do we need to be doing differently to avoid this?
I think AI bias is a very important issue. More broadly put, it's the black box problem. Based on technology like deep learning, we can have an AI agent that has absorbed a great deal of data, and created elaborate decision paths- and then outputs conclusions and we don't really know where they came from. What exactly was behind these outcomes? It is the classic black box. This is, I think, is what's pushing a lot of work now to create insight into the process of cognitive computing. In highly regulated industries, where gender bias, racial bias, things that are literally illegal, you need to know if your machine is inadvertently making biased decisions.
Fortunately, people are seeing this as an opportunity to create systems that give us visibility into the black box and explain in natural language what's going on. There's already some pretty impressive work out there. The first line that's driving it, I think, is the regulated industries. But past that, ultimately you may end up changing fundamental business strategy based on what an AI tells you to do. There's a huge responsibility to understand what's behind that reasoning.
So, to summarize, yes, this is a crucial question and problem. Fortunately, it's been recognized and I think a lot of good work is being done to start addressing AI bias.
Why excites you the most about intelligent automation, AI, cognitive, etc.? How do you think these innovations will change the world for the better?
First of all, what has made humans successful as a species is really our intelligence. That's our secret sauce and it's gotten us to an incredible position on this planet. We've also not been smart in some ways, so I think some additional intelligence can always be of assistance.
Secondly, AI, cognitive computing, etc, is coming at such a great moment. Because we are also living in a time where we are accumulating enormous new amounts of data – through everything from GPS to smart sensors to mapping satellites - that provides perfect fodder for the machines of AI and cognitive computing.
So if we did not have AI, we would be awash in data and not really know what to do with it because we aren't quite that smart. This is where cognitive computing and AI will really fit in. We're a smart species, that's how we got where we are today, but we in order to keep going in a constructive and sustainable way, we need to get smarter. Intelligent technology is a terrific way for us to get much, much smarter.
*This Q&A was
originally promoted for the 2018 Intelligent Automation Week.