Explainability etc. New Audit Trails for Exponential Technology

Fashionable Exponential Technologies

There are always a handful of ‘fashionable exponential’ technologies being touted as world changing. As I write we have as a sample Artificial Intelligence (AI), blockchain, autonomous vehicles, renewable energy. Notable about them is that they push out easy human understanding of their function and maintenance. Importantly though, in the case of a significant event involving them, is our ability to understand the rules and events that led to it. What does that mean to potential users of them in business?

Evolution of Technologies

Technologies develop with the intention of an improved lot for people. They do however go through cycles of improvement. In those cycles they iterate against human usage and find their way (things needed, things to be discarded and things to be fixed), change to keep pace with technological improvement and, they need to find ways to integrate with governance and industry practice. In the case of the latter let’s consider one new technology, Artificial Intelligence.

Consider Consumers

Decision making has advanced from paper borne policies to rules expressed in computer systems. For consumer payments approved by computers for example, networks apply database checks, security authentication and commercial rules around things such as credit limits. This is becoming increasingly complex to account for context issues such as location, velocity and particular customer behaviour. While there are efficiencies in electronic transactions dispute handing is a costly exercise with this complexity.

In the case of a consumer payment transaction dispute, historically a clear understanding of the data and rules applied can be found. The advent of Artificial Intelligence makes the decision making potentially less clear.

Thoughtful Disputes and their Explainability

Artificial Intelligence is not just a case of more complex rules. It is the application of basic rules but also the inference of patterns from historic, real-time data and context. As a result, it can provide more sophisticated decision making. When AI is used in consumer facing business, human understandable reasoning behind decisions must be available to deal with customer disputes. In the case of AI, there are moves to encourage what is called ‘Explainable AI’ that do this.

Why is this useful for business?

When considering the integration and capability of any new technology more than the step change in function needs to be considered. The ability to handle disputes and to improve facilities through edge-case incidents also need to be integrated. During transition multiple systems may be used. This applies as much to AI as to to using blockchain for asset management across borders and hence governance, autonomous cars involved in accidents and renewable energy source outages.  In the case of Artifical Intelligence, the dispute handling may be less human intensive as a result of the technology itself providing it is crafted/trained well. It is a meta-technology, one may be able to explain and justifiy its behaviour not just to its owners but its users.

Interested in more thoughts like this?

If you found this interesting and worth exploration with your organisation let’s talk about how we might work together.  To find out how to get in touch go here

Interested in what I do but the timing is not right yet to do something together? I regularly publish some of the insights I use, and thoughts that I have in a briefing. You can sign up for that below

I’d like to hear from you.

Subscribe to the Future View

A regular, short and advert free briefing on the latest ideas from my work. If you like a different and deeper view please sign up. I respect your privacy