Empathy in AI
In a recent podcast, Sam Charrington and Rob Walker discuss the implications and role of empathetic AI systems in decision making.
The conversation is mainly about whether it is possible that in the customer strategy, empathy is well represented or tangible?
When building empathetic AI systems, the important question is whether the next action the machine learning system takes is the right or optimal one for a specific customer given a set of metrics or factors. At the end of the day, the goal is for the customer to obtain some value from an interaction with the AI system — and that’s where empathy comes into play.
A pro-empathy strategy ensures that the customer is always considered in the decision-making process. Companies are always seeking the next best decision given some risk or context, also referred to as the “next best action” capability.
Rob says that it’s a good idea to have companies monitor the effects that their technology’s decision-making processes are having on customers — a dashboard tool of some sort. The goal is for a company to always keep high moral standards and ensure that the customer is not in a losing position, and to always seek to generate some value through services. In addition, the empathetic AI system should always seek to offer the customer a service or option that they are pleased with — also referred to as relevance. This can be learned automatically through the historical interactions that the systems have with customers and the customers’ responses.
Overall, Rob advocates for companies to consider incorporating empathy metrics in AI algorithms when making the decision about the kinds of services/options to offer the customer. In addition, he also hints that other important profitability metrics and more rule-based generated insights can also be parallelized into the system’s decision-making process.
Specifically, some of the patterns the system should consider when taking the “best next action” are suitability, risk mitigation, context awareness, profitability, mutual value, and relevancy. These are all factors that should be embedded in the decision-making framework to ensure higher transparency.
This helps brands/companies to avoid misselling, while encouraging best practices and properly managing ethics/bias issues revolving AI systems.
One great point that Rob makes towards the end of the discussion is that AI systems should be carefully designed to practice empathy because it helps companies to establish better long-term relationships with customers. He adds that empathy is a need that evolved from a desire to collaborate.
