If you believe Hollywood and the papers, the two major fears about artificial intelligence (AI) are: “AI will put me out of a job!” and “The robots will kill us all!”
As much fun as it might be to write science fiction, I am not going to write about killer robots today. And while the first fear is perhaps a real one for less flexible workers, I think there’s a lot more promise than threat in the current AI forecast. If we look at the example of how automation transformed manufacturing, there was first a real golden age of economic growth. It was only when the entire sector lost momentum that there were problems for people and machines working together. History will probably repeat itself as automation and AI move into the knowledge economy: before any large-scale white-collar job losses, we can expect a boom that helps everyone.
How did manufacturing automation reach that historical sweet spot where it enhanced productivity without cannibalizing the industry? The technology matured through a process of iterative improvement. Occasionally there’s a flaw that no one has the will to fix, and regulation can step in. (For example, Europe has a lot of regulations regarding safe escalator design; the USA and China have fewer, and continue to have many escalator-related injuries and fatalities.)
Just in recent years, tech automation and AI have reached the same point. Iteration after iteration has worked the bugs out. A platform like Katabat’s, which can handle the compatibility quirks of an institution’s legacy systems, is excellently placed to take charge and start quietly orchestrating automated tasks in the background.
Why or Why Not?
Businesses today are eager to invest in AI and machine learning because of the potential for amazing growth. But business leaders need to take a step back and realize that each individual enterprise has its own needs and priorities, which differ from the market as a whole. For many, a chief concern is cost. AI hiring is vastly outpacing the supply of skilled researchers and programmers. The big Silicon Valley firms pick up huge numbers of talented individuals at enormous salaries, but companies with more limited resources are left unable to compete.
Other organizations have plenty of capital, but are constrained by other factors such as regulation. For large global banks like our clients, it is imperative to be able to demonstrate compliance. Automation presents a great opportunity to remove human error from the equation. However, AI and machine learning, if implemented and trained imperfectly, could exacerbate human problems or innovate new negative or non-compliant behaviors. The ill-advised experiment of Microsoft’s Tay chatbot, which learned to tweet like a racist in just a day, is a cautionary tale. Consumer lending has a long-standing problem of bias. It would be easy for this bias, for instance, to make its way into AI training datasets, with the result of a lot of regulatory action and bad press for banks.
The Right Solution at the Right Time
Katabat specializes in automated, intelligence-amplifying solutions because they are mature technology. Well-written automated workflows add immense value without a downside. You basically need to retain one strategist competent to adjust the workflows, and you are set! No longer does every single agent have to keep abreast of all the regulations. We have worked for years to facilitate adoption of new tech by big finance firms. Sometimes it pushes them outside their comfort zone, but we don’t put them in a position where the rate of change is destabilizing for their customers.
In my next post, I’ll look at how we judge when it is the right time to supply our customers with a new technology, such as expanding AI into new areas of their business.
You can always contact me at email@example.com with any thoughts or questions on AI and automation.