It sounds paradoxical: can something artificial help us understand what's real and what isn't? Social media platforms like Facebook and Twitter are pinning their hopes on this possibility in their fight against "fake news."
Can't Spell Artificial Intelligence Without Art
Recently, I had the good fortune to visit an exhibition in Paris entitled "Artists and Robots" with my wife, a former art teacher. As a former artificial intelligence (AI) researcher, it’s not often we get to attend an exhibition that can appeal to both artists and scientists. Art is not my strong suit, but I do understand it’s not meant to display "reality," but rather, as my good wife instructed me, the artist’s “interpretation of reality, designed to appeal to his or her audience.”
What immediately became apparent upon entering the gallery was the exhibitors were not geeks wanting to become artists, but artists who wanted to augment their capabilities with AI algorithms. Note the word "augment," not "replace." What’s more, the AI used was not the convergent problem-solving algorithms we are used to seeing, but the emergent ones that produce unpredictable results (computer coders call these bugs). I’m no expert on modern art, but I suspect something like the following image would pass the Turing test of not being able to tell if it was produced by a human or a machine:
Perhaps the most stunning exhibit for me, being the scientist struggling to understand art, was how Spanish artist Joan Fontcuberta was able to use sophisticated simulation software, designed to create 3-D landscape visualizations from mapping coordinates, to create landscapes (image on the right) not from mapping co-ordinates, but from signals drawn from artworks like the one from post-impressionist painter Paul Cezanne, shown on the left.
Having also recently visited Cezanne’s studio in Provence and seeing the actual landscapes he painted, I can testify Fontcuberta has done a stunning job of recreating reality from the artist’s interpretation. Has Fontcuberta stumbled upon a way to identify the gap between a qualitative interpretation (e.g. fake news) from reality?
Perhaps this is a bit of a stretch, but in essence this is what we want AI to do for us. Can AI provide a commentary on what we are continuously exposed to in our digital environments and what can be objectively assessed as reality? With the advent of digital photography and applications like Photoshop making it easy to manipulate supposedly "real" images to attract a targeted audience, we need all the help we can get.
AI has traditionally been good at the quantitative aspects of problem solving, reducing seemingly quite complex problems to simpler quantitative decisions, largely through the weight of processing millions of sample cases to compare and learn from. But what about the more qualitative aspects of business, such as leadership? Aren’t we regularly told that leadership can be both an art and a science?
Learning to Lead With Facts
A recent McKinsey article postulated AI can be used to help people become better leaders. The article suggested by using AI analytics to predict and/or propose potential future states, leaders can potentially become more agile and therefore more effective. Substantially, the suggested AI analytics were people-centered and more akin to the social network analytics (SNA) companies like SWOOP and others develop, which help leaders see the real collaboration patterns happening across their organizations. While not referred to as AI, SNA does share the same graph-based search structures often employed in AI, and therefore provide a convenient complement to AI in supporting leaders.
As a leader in an organization, one is often faced with ambiguous signals from the shop floor as to what is really going on. One former CEO of a major telecommunications company once told us hierarchy can really get in the way of senior executives having authentic conversations with staff, with information flows both up and down the hierarchy being carefully orchestrated. The result is you only hear what people want you to hear. So how can senior leaders separate fact from fiction?
Continuing the theme from the McKinsey article, we would argue that social network analytics can provide much of what leaders require. Having conducted over 100 Social Network Analysis (SNA) for organizations of all sizes over the past decade or more, one of the most common comments we get from senior leaders when they see the social network maps (sociograms) we produce, is how clearly they reflect what is really going on in their respective organizations. The sociograms appears visually qualitative, but each node and link can be traced to one or more data points. In this way leaders can gain a qualitative ‘big picture’ view, or choose to drill right down into the detail of individual relationships.
Peeling Back the Data LayersI started this article with a story about art, so it seems appropriate to finish on that note.
Some years ago I was visiting fellow social network analyst Valdis Krebs in his hometown of Cleveland, when he suggested we go to an art exhibition. This was no ordinary art exhibition, but one featuring the work of Mark Lombardi, whose hand-drawn sociograms had attracted the attention of the art world. The subject of his social network analysis was global crime networks and money laundering. In fact, after appreciating the artistic elements of his pictures, one could drill deeper to explore the data contained within his art. Apparently a number of the early viewers of his exhibited art were from the FBI.
So can something as artificial as artificial intelligence really help us separate fact from fiction? Real news from fake news? The short answer is that more than ever we are going to have to check, cross check, critically review with trusted partners and more, if we are to reliably separate fact from fiction in our digitally immersed workplaces. And AI will definitely have a part to play.