Corporations May Need a Chief Artificial Intelligence Officer, says "AI Superpowers" Author Kai-Fu Lee
Former Google China President and best-selling AI author Kai-Fu Lee provides a board wake up call, and advice
State Farm’s recent ad campaign mocking insurance companies using artificial intelligence (AI) includes a clumsy and ineffectual robot attempting to provide service to two pro basketball stars trying to file a claim for a dented car.
In explaining the robot to the basketball players, the human agent says: “The budget insurance companies are building these cheap knock off robots to compete with us.”
The spot quickly got panned on social media, and one of their AI-focused competitors even paid to promote the ad as a way to promote their own business.
It’s a battle now playing out in almost every industry, between companies that may still see AI as all hype and AI startups and traditional corporations that have embraced AI, says Kai-Fu Lee, author of AI Superpowers: China, Silicon Valley and the New World Order, and the former president of Google China who is now a venture capitalist based in Beijing and a board member of electronics manufacturer Foxconn Technology Group.
Lee sees State Farm’s ad as an “extreme example of foolish behavior by a company” because there are parts of the insurance giant’s business, and almost any type of business, that can and will be replaced by AI.
Companies and directors, he adds, “shouldn’t put their heads in the sand. They need to understand the full impact of AI.”
AI is already here, and it’s not about inept robots or robot overlords. It’s about data — data used to predict shopping patterns, screen job applicants; move autonomous cars, power Siri and Alexa, and accelerate the insurance claims process through automation, just to name a few.
Boards need to be asking management critical questions now, Lee stresses, including whether a start up is doing what your company is doing. “You have a legacy, and all the people want to extend the legacy rather than disrupt a stable situation,” he says about companies that aren’t fully embracing AI. “The disruptors from the outside have nothing to lose and they will do your business for one tenth of the profit because they’re automated and leaner.”
How will AI impact business?
An excerpt from Lee’s book sums up the potential well:
Insurance companies have been covering accidents and catching fraud, banks have been issuing loans and documenting repayment rates, and hospitals have been keeping records of diagnoses and survival rates. All of these actions generate labeled data points — a set of characteristics and a meaningful outcome — but until recently, most traditional businesses had a hard time exploiting that data for better results.
Business AI mines these databases for hidden correlations that often escape the naked eye and human brain. It draws on all the historic decisions and outcomes within an organization and uses labeled data to train an algorithm that can outperform even the most experienced human practitioners. That’s because humans normally make predictions on the basis of strong features, a handful of data points that are highly correlated to a specific outcome, often in a clear cause-and-effect relationship.
For example, in predicting the likelihood of someone contracting diabetes, a person’s weight and body mass index are strong features. AI algorithms do indeed factor in these strong features, but they also look at thousands of other weak features: peripheral data points that might appear unrelated to the outcome but contain some predictive power when combined across tens of millions of examples. These subtle correlations are often impossible for any human to explain in terms of cause and effect: why do borrowers who take out loans on a Wednesday repay those loans faster? But algorithms that can combine thousands of those weak and strong features — often using complex mathematical relationships indecipherable to a human brain — will outperform even top-notch humans at many analytical business tasks.
So where should directors even start when it comes to uncovering the promises of AI? Lee sat down for a Skype interview with Directors & Boards to provide his take on what issues should be on the boardroom agenda now.
Who should be the AI gatekeeper — the CIO, the CTO or some totally new person, maybe a Chief Artificial Intelligence Officer? Should there be a CAIO?
It depends on the type of the company. Let’s say you’re an Amazon, that’s probably the entire structure perfect. You don’t need to hire anyone. You know AI and you have a large workforce, and all the implications are well understood. Basically the CEO and probably to a much lesser extent, the CTO would just simply work together.
For traditional companies, it’s kind of tough. Unless they have a very sophisticated CTO who’s knowledgeable about the multi facets of AI, how to use it to make money, how to use it to save money, you might need a CAIO. I personally don’t think CAIO is a long-lasting position, but if you appoint someone to do that for a few years I think that’s reasonable and see how it goes. Just like there’s no chief internet officer because AI is supposed to become easier to use, and more and more understood, and go down to the manager level probably within three years. In the early days of internet there were companies that had chief information officers, or internet divisions, knowing or not knowing they’d go away.
They need an expert to help them navigate. That could be a CTO or whoever on the staff who’s knowledgeable about the tech issues, or hiring a consulting company either a McKinsey, or KPMG, that type of company. They could at least help you map issues if you don’t have someone in your company who can take that role, or hire a consultant to do that. That’s probably the beginning point to trot out the overall blueprint of opportunities and issues.
How can they ensure the CEO is keeping on top of AI advances that could bolster business or pose a competitive threat? How do you know if your CEO is watching the AI store?
By presenting a plan that the board can approve. Ask for an AI plan. Maybe invite an AI expert to the board meeting to help review the plan. That expert could be an industry person, someone who’s done AI implementations. Maybe a start up entrepreneur who wants to work with the company or a VC who has a good relationship, or someone who’s been with an AI company, or a management consultant with the right people
If it’s critical enough, they can add a board member who’s knowledgeable in this space, I’ve certainly gotten a lot of invitations, but I’m not able to take on any more. But it seems to be on people’s minds.
Outside my normal investing scope, I am on three boards, all related to AI in either significant way, or somewhat significant way. These are multi-billion dollar companies that I didn’t invest in. I agreed to perform the board membership because I thought those companies could be of value to my VC business and they wanted my AI expertise. The $10-billion-plus companies can find someone like me, for whom it’s a win-win situation.
I’m a board member of Foxconn. I’m helping with injection of AI in manufacturing. They basically presented divisional-level AI plans, which the CEO, chairman cares greatly about and I would be there to critique them.
In some cases, I was able to bring our portfolio companies to help them. In other cases, I recommended industry experts to them who knew more than I did on a particular problem. They formed several joint ventures. They publicly announced seven or eight investments that are related to advancing the roadmap in AI in some way. To improve their corporate process, to improve the quality of the products, assist with autonomous inspections, look at instrumentation of the manufacturing process, and looking at the automation roadmap of the company, and also investing in AI hardware companies. I have not gotten into the people-reduction side. There business is both growing and complex. It’s not poised for AI displacement in the short term.
What questions should directors be asking?
First, on business side, I would ask are any of your processes automatable over the next five years, and are there startups looking to automate your process. If you ask the company’s business managers, they’ll always say “no.” For them to take a big risk in using AI and letting go of half the workforce, it’s a big risk, why should they take it, they’re just career managers.
Another question is, also consider competitors. When you understand the true impact of AI over a five-year horizon, it’s not a question of do we choose to do it or not do it here are the HR implications, and cost implications, which is better for us? You can’t evaluate your company in a vacuum. If you don’t cannibalize yourself someone else will, either a start up company with nothing to lose, no baggage, no legacy, or another traditional company that chooses to adopt it.
Adopting AI for the first two years might not be very good ROI [return on investment], because you still have to keep all the people, and you have to pay for the technology and test if it works. You can easily have a mid-level manager come up with a case of saying, “not a good deal, let’s wait”. But by waiting it means somebody else may take a two-year hit on the P&L but in three or five years end up with something dramatically better than you and then eat your lunch.
So you’ve got to the five-year outlook. I think five year is about the right timeframe because if can’t be done in five years then it’s probably too soon to worry about. For many companies it may be too soon.
What if an organization is at the early stages? What is the critical beginning point?
Companies have to build their own data. That data may mean sales data, customer data, usage data, website usage data. If your data is opaque you’ll have to find a way to redirect it to a more transparent method.
For example, if you currently have a giant team that calls your customers and picks out the likely ones for sales leads, that needs to be recorded. Or maybe some of the interaction needs to move online.
Owning the destiny of your data, and using online and transparent methods, and using good data-warehousing techniques is the premise of doing AI. Otherwise, AI can’t just come in and replace a data-less company. If you’re a data-less company, then there’s nothing AI can do for you.
Kai-Fu Lee is chairman and CEO of Sinovation Ventures, a venture capital firm managing $1.7 billion in investment funds focused on developing the next generation of Chinese high-tech companies. Prior to founding Sinovation in 2009, he was the president of Google China, and previously held executive positions at Microsoft, SGI, and Apple. He received his Bachelor degree from Computer Science from Columbia University, Ph.D. from Carnegie Mellon University. In the field of artificial intelligence, Lee founded Microsoft Research China, which was named as the hottest research lab by MIT Technology Review. Later renamed Microsoft Research Asia, this institute trained the great majority of AI leaders in China, including CTOs or AI heads at Baidu, Tencent, Alibaba, Lenovo, Huawei, and Haier. At Apple, Lee led AI projects in speech and natural language.