Pega's Rob Walker predicted exactly that earlier during a keynote address this year at Pega's global customer event, Pegaworld in Las Vegas.
How many ‘likes’ does AI need to get to "know" you?
- Just 10 to know you better than your colleagues
- 70 to know you better than your friends
- 300 to know you better than your parents or spouse
- 700 to know you better than you know yourself
Understanding the difference between Opaque and Transparent AI is crucial for the Insurance industry.
Transparent can be audited and is capable of explaining why and how it reached certain conclusions.
Opaque cannot be audited nor give reasoned behavior - Microsoft’s had a fiasco with the opaque Tay Twitter chatbot, that started issuing racist, sexist, and trolling messages.
Both insurers and intermediaries need to examine what form of AI can work for their businesses and the value that it can create.
With the continued blurring of lines with insurers/brokers and the use of artificial intelligence in financial services, needless to say the industry should be wary of the risks as they exist, along with the potential benefits. Among the threats, according to the Financial Stability Board (FSB), is the possible rise of third-party dependencies – and what this could bring about. We could be setting ourselves up to rely on players that are not bound by regulations. The FSB also noted the lack of interpretability or auditability of machine learning methods, thus posing a risk at a macro level.