512.123.4567 sales@malauzai.com

Chatbots will be smart but artificial intelligence doesn't exist yet.

By Robb Gaynor

Chatbots will be smart, well-informed even but not artificially intelligent. Computers cannot think by themselves. They don’t self-correct. They don’t learn on their own. No! Artificial Intelligence does not exist, and all this talk about how financial chatbots will be driven by AI is hogwash, not true, doesn’t exist and it won’t help chatbots, whatever those are. Yes, a goal is to achieve artificial intelligence. Think: a computer that thinks on its own, that is sentient and learns, synthesizes thought. My rudimentary understanding is that in fact, we are far, far away from this. Someday yes, but not today and surely not in the digital banking channels, let’s get real.

We are about as far from a self-generating, self-correcting environment as you can get. At Malauzai we have 10 releases a year and our clients absorb about 2 of those. There is lots of stuff that goes into having the digital banking channel “learn” automagically. Like writing requirements, coding, testing, acceptance testing etc. Imagine going from that world to one where the actual program changes itself. It senses that a new feature is needed. It builds it, front to back. Tests, and then rolls it out, all in some automated, self-generating fashion. Hahaha. That’s funny. OK, it will happen. Programs, software does learn and self-correct today. But not substantially. So there goes self-learning and self-generating.

We can however fake it, we can programmatically make it look like our programs are intelligent. Now I’m not saying we can programmatically learn. But it is self-determined. We program in our intent. To whatever extent we can make the program flexible and attached to data, it might emulate learning. Take dynamic personalization. Common now is a feature in digital where the first screen you land on is based on the screen you use the most. It’s dynamic and based on usage data. Whatever feature you used most (in this case most is defined as 7 out of 10 times) shows up the next time you log in. It’s smart. It uses data to drive personalization. And it is dynamic in that it changes. But it’s not AI. We use the 7 out of 10 rule. We programmed in the dynamism to the front end to be able to show the right screen. I will give you this, it approximates intelligence. And it is way, way more dynamic than a static digital banking session. But it is limited in what it can do and a human had to give it the parameters within which it functions. This is what we can do today. We can approximate. And the more data we have, the better we are at it. But it requires analysis. And programmatic change. Slow. And not AI.

AI will come. And I am being extra hard on what we are doing today, as it does approximate intelligence even if it is deterministic (my made up word, not sure it is used in the right context) and programmatic. Fortunately in our world of financial services, there are only so many products, only so many things a person can do. And we have studied for years their usage patterns. So in some ways we can approximate intelligence to a greater degree than some other industries. Financial services might just have an edge as it attempts to make its chatbots smart, sentient. Chatbots by the way will need even the limited intelligence we can program in. Voice and natural language, which are at the heart of chatbot technology, requires a high level of dynamism. The bots need to appear natural, thus rudimentary techniques to emulate AI will be important. Knowing the data, knowing intent will be critical. AI might not be here today, but it will come and FS will be ready. My banks chatbot will be smart from the get go, or so they will appear! Long live fake AI.

Retraction: Due to overwhelming feedback about my statement that AI does not exist, will never exist and is basically a fantasy, I fully retract the statement that AI is dead or does not exist. That was fake news. Sorry. Read the next blog for a full and earnest disclosure of the retraction and all associated logic and supporting rhetoric.

Robb Gaynor

Robb Gaynor

Founder & Chief Product Officer

Robb heads up the product development and marketing of the organization, focused on ensuring the company has solutions at the cutting-edge of mobile innovation. Robb has over 23 years of experience in the financial services industry, having previously served in leadership roles at Union Bank of California, Digital Insight/Intuit, and other large global financial organizations including Swiss Bank Corp, Charles Schwab and Wells Fargo Bank.