Recently I had a friend lose his 42-year-old wife to breast cancer. The entire painful ordeal was played out on Facebook: they both talked about her treatments, her ups and downs, her eventual tragic decline. Thousands of comments were made on the postings over the course of 2016, the year she passed away, At year’s end, Facebook created a retrospective video of all the media and posts, and played it all back to my friend, assuming (due to an algorithm that relied purely on numbers) that he wanted to relive it all. This, in spite of the fact that my friend had memorialised his wife’s page, and sent Facebook her death certificate. “All that information I gave them,” he lamented, “and they handled it in a tone-deaf manner. Facebook has no tact.” Last week I gave a keynote on this topic in Denmark and after telling the Facebook story, the phrase came serendipitously out of my mouth: “That’s not artificial intelligence, it’s artificial stupidity.”
Software is very good at managing, finding, and distributing information, but so far it is still lacking humanity. This, primarily, is why so many people lack trust in technology: it does not understand or react properly to how we feel. Current consumer and enterprise technology does not respond differently based on our tone of voice, it cannot identify when we are exasperated, or happy. Technology lacks empathy. As we think about how technology and humans can work together more symbiotically, as we look towards a future where we want both our work and personal lives to retain what makes them worth living – technology will need to better understand, respond to, and adapt to human feelings.
Software makes decisions on what to do based on lines of code and data inputs. Often the decisions are based on quantity, rather than quality or discernment — which is why Facebook thinks it’s a good idea to play back the drama of someone’s death. High quantity of reactions = important, with no consideration of happy vs. sad. But if Facebook had considered the actual language expressed, the “sad” icons clicked, an algorithm could have quite easily ascertained that this was not a happy occasion.
“Software is very good at managing, finding and distributing information, but so far it is still lacking empathy.”
We are only just starting to model human sentiment and incorporate it into data structures that software can actually use. This is being driven by technologists and business people alike who slowly realise they can use customer sentiment to offer better emotional experiences. This imaginative leap is not just about traditional online content publishing and marketing, which uses customer data to target content. This is about using technology to layer our communications based on human reactions and empathy, as if it were a human-to-human relationship.
To truly improve how we interact with technology, we need to rely on both people and artificial intelligence (AI) to better interpret response, and then in turn embed empathy-oriented metadata into digital assets (pictures, video). This will enable more specific forms of content targeting and reuse that are tuned into people’s emotions.
Empathy is what our data lacks — currently we are fed content based on demographics and what employers or marketers think will work, rather than real reactions and feelings. However we have never had more opportunities to give, and better use, human feedback. We are becoming quite skilled at gathering human sentiment, but most companies, like Facebook, react to the intelligence in a tone deaf manner. To bridge the gap, we need to properly analyse reactions as much as the data itself, feed it back into the software, and then re-use optimally. Essentially we need to teach software to understand and react based on data about emotions. This will make technology, and our future, far more human.
Theresa Regli is a partner at Jericho Chambers and Chief Strategy Officer for KlarisIP. She has advised hundreds of businesses on data design and content technology and is author of the definitive book on managing digital marketing & media assets Digital & Marketing Asset Management: The Real Story of DAM Technology and Practice.