As we continue to advance the capabilities of AI and its use in various applications to make the world a better place, the debate is emerging around whether we can or should infuse these systems with what is considered uniquely human emotions, such as empathy.  

On the second day of NASSCOM Xperience AI Virtual Summit 2022, Alan Trefler, Founder and CEO of Pegasystems, delivered the closing Keynote on the role of empathy and emotional intelligence in designing inclusive technologies and deconstructing the biases encoded into AI.  

AI should be trustworthy to make it work the way businesses and enterprises want. Mentioned following are some of the elements that constitute reliable AI: 

• It should not feel artificial 
• It should be actionable 
• It should augment humans as individuals 
• It should be aiding us continually 

Being smart is technology's biggest strength, but it is sometimes insufficient. "AI should be able to decide the best actions", says Trefler. It should be able to facilitate the execution of a particular action so that humans can operationalize them.  

"Just knowing action and being able to make it is not useful. It should be able to judge and measure the outcome", he added. 

Ethical Bias Check 

When a machine is being trained, human beings feed data into it. Therefore, there are chances that the machines may take constructive and non-empathetic decisions under human influence. This will affect the overall development of empathetic AI. 

 Pega's Ethical Bias Check ensures that no unintentional bias is hidden in your next-best-action strategies, either within the models or in the business. Users can define fields with 'bias potential'- like age, ethnicity, gender or income- then simulate the strategies where it is ensured that they are not skewed unfairly towards or away from specific groups.  

"Ethical Bias Check identifies the biases that we know exist but do not want in our system", says Alan Trefler. 

Here is how it functions: 

• Define the fields to monitor for bias 
• Simulate strategies and automatically test 
• Generate warnings when bias issues are found 
• Identify the problematic actions, issues and groups 
• Adjust those strategies, including logic, models, rules etc., to validate or resolve the issue 

Customer Engagement Engine 

Most customers don't believe the organizations they deal with. "We need to go above and beyond because there are many things that need to be distilled to meet the needs of the customer", says Trefler.  

Companies need to be able to sense what is going on. Tasks must be completed in real-time to have effective AI. The AI must be interpretive of the customer's needs in specific business contexts.  

The Customer Engagement Engine can bring together the factors that are important for customers. Alan Trefler opines, "It enables the business to figure out what is important to their client in that point of time". Being able to complete such pragmatic tasks is the actual application of AI.  

Working Smart 

The idea of a self-optimizing company is the successful future of AI. Three significant factors ensure that a company is working smartly: 

• Making the right decision 
• Executing the decision 
• Recording and learning from results 

 Decision-making and execution are an ongoing cycle. It aids in the process of learning and growing. Also, organizations must think about equity and empathy in what they do.