Future AI: The rise of empathetic machines
An excerpt from eConsultancy’s 2019 “Digital Outlook” where Beyond’s Jenai Marinkovic and Kim Turley, talk about the future of AI and the empathy gap can be seen below.
From the smartphones we carry, social media we follow, music we stream and navigation apps we use, to the voice assistants and smart home devices we welcome into our apartments, Artificial Intelligence is already proving its value in our everyday lives.
Today’s AI excels at tasks which require hard skills: gathering and synthesizing straightforward information to make decisions. It is accurate when analyzing data to report results, whether that is traffic on the roads to help a steady flow of vehicles, or market rises and falls to make trading decisions.
While it involves upfront investment, AI does not expect a break from work — unlike its human counterparts — and produces results 24 hours a day, which reduces costs and increases productivity. And while to err is human, to be consistent is within the fabric of AI.
With its potential to make trading decisions at speeds of which humans are simply incapable and its ability to interpret medical images more accurately, in the short-term it is already starting to replace hard-skill jobs such as finance analysts and radiologists.
But this is just the tip of the AI iceberg. We’re still on the cusp of the fourth Industrial Revolution and when AI realises its full potential we’ll see widespread changes across almost all industries.
Currently, as AI is more effective at particular tasks, rather than full workflows, it can be frustrating for users: just think how many people end up losing patience with Cortana and Alexa.
The empathy gap
The biggest limitation is that AI cannot yet convincingly go beyond anticipating people’s needs, to also show empathy. It’s this struggle for machines to operate with emotional intelligence that has held AI back so far. In times of trouble, confiding in your AI assistant is unlikely to bring comfort.
Robotics researchers at Massachusetts Institute of Technology have found that humans interact better with robots that respond to emotions on our faces. But teaching machines soft-skills, such as responding to feelings and communicating with nuance, is a huge challenge. The main limitation comes from humans themselves: we don’t have enough understanding of diversity in human mindsets and lack self-awareness of our own vulnerabilities. This makes it difficult — and in some cases impossible — for AI’s teachers to convey to machines how people really think and feel.
The industry is showing a growing awareness of the need to develop Artificial Empathy, or ‘AE’. But most of the research surrounds teaching AI services to be empathetic, without being mindful that AI’s teachers need more sensitivity and understanding of human culture and emotional need, before machines can echo them. In future, we are likely to see more collaboration between specialists such as psychologists and technology practitioners to teach a deeper understanding of how humans feel.
What is coming next
Over the next two years, we will continue to see AI take over tasks that involve analyzing data to report results. But there will be increasing connections between AI and humans, with chatbots used throughout the service industry, helping people quickly reserve restaurant tables, navigate forms, or acting as digital personal assistants.
AI is already being used to learn from a company’s best performing workers, and teach others to emulate their success by Cresta, a company run by a team from Stanford. In years to come, the use of machines to improve performance in this way will become commonplace.
One of the most exciting developments that we expect to see in the longer term is a series of breakthroughs that give AI better emotional intelligence, which will transform jobs and the workforce. This is something to which future-facing companies are already devoting their efforts, and technology leaders in communication, language and understanding of human intention will lead this second wave in AI.
It places the emphasis on humans to increase our understanding of ourselves: as we grow in empathy, so will machines and AI will realise its full potential. Compassion has never been more important — and that is a great thing for society.
Once machines start understanding the nuance of human interaction, and that ‘I’m fine’ can mean anything from ‘I’m well’ to ‘I’m furious’, they will be able to offer meaningful support to care workers, therapists, even marriage guidance counselors.