Our trip to the AI: More Than Human exhibition
Always looking for ways to learn more about how technology and how AI is impacting our lives, the team at Pure360 headed off to the AI: More than Human exhibition currently running at the Barbican in London.
In this blog, a few of our experts give us insight into what learnings they took away from the exhibition and share their thoughts on how AI will be further integrated into our lives.
To me the concept of Artificial Intelligence felt rooted in modern society, however after attending the exhibition I was blown away by the fact that humans have always been fascinated by the artificial creation of living beings as well as finding ways in which to extend the mind. This is the root of AI, which dates back to the times of the Bible and possibly even before (it is just wasn’t well documented).
One quote stood out for me “What are the computers and robots of our time if not golems” by Isaac Bashevis Singer. So what is a Golem if it isn’t just the character created by Marvel?! It comes from Jewish mystical tradition where creating a ‘Golem’ was understood as a means to grow closer to the Creator and to achieve spiritual perfection. And from then the Golem became the subject of legend, folklore and, fast forward many hundreds of years, the subject of numerous literary texts. The Golem characters in literary and visual art can be seen in works such as Mary Shelley’s Frankenstein, the painting “Dream of a Golem” by Jules Kirschenbaum to the more modern Blade Runner and ex Machina.
The existence of such a creation and humankind’s strive for knowledge and mastery has also led to the question of what it is to be human and consciousness. Not only that it questions the role of humans playing God, the relationship between masters and slaves.. but what happens when these creations get out of control and what are the consequences of self learning? From these fears come our concerns over apocalyptic AI. As our creations have turned into intelligent machines are we heading towards a world where we are no longer able to differentiate, where the machines take over our jobs and our world?”
The exhibition, although not leading with a marketing viewpoint, demonstrated some of the core AI technologies being used in marketing today. Natural Language generation and processing and their use in chat bots. Image recognition and the process of training a model to recognise stages of a tulips development. Citizen profiling in China to enable their social credit scoring that drives access to schools, use of transport and loans to name just a few things.
All of these technologies are now readily found in marketing AI systems. However our willingness to interact with these technologies as consumers is limited by our trust. Hollywood has concentrated on what can go wrong with AI over the benefits if we get it right and hence our trust in anything AI is very low. In order to build this trust we need to build greater understanding of what AI is and be more transparent data. Businesses must enable us to see and interact with the data held on us and teach us how the technology works if they want to reap the benefits.
Computer vision has many applications from detecting manufacturing defects to helping self-driving cars see and interpret the world around them. But how computers see is different to how we see and this leads to some surprising effects.
I accidentally managed to carry out an ‘adversarial attack’ on one of the exhibits, an interactive driving simulator that analyses your facial expressions to determine your emotional state as you drive. By chewing gum as I used the driving simulator, my facial expression was analysed as angry when in fact I was smiling. This is of course an innocuous example but consider this applied to the realm of self driving cars. Could the vandalism of traffic signs or road markings cause accidents? Much like the co-evolution of virus creation and anti-virus software, it seems that AI will need to develop its own safeguards to combat the threat of adversarial attacks.
It came as no surprise to me that I left the exhibition still believing that AI is a double-edge sword that needs to be handled with care.
In one moment, I was listening to an expert from MIT explain how AI was helping predict breast cancer 5 years in advance of it developing. From looking at thousands of mammograms from patients who had already developed cancer within five years, the machine has learnt to pick up on the most subtle of changes within the pattern of the breast tissue to predict if cancer will develop – something the human eye couldn’t possibly pick up on. This presents the opportunity to provide personalised care to individuals, preventing unnecessary and invasive procedures such as mastectomies as well as saving many more lives.
However, almost immediately next to this source of inspiring news was a rather horrifying breakdown of how easy it was to literally put words into someone’s mouth using AI. The University of Washington used neural networks to mimic Barack Obama’s speech patterns and “mouth shape” to recreate a brand new speech. The results were chilling, as I watched Barack Obama present an entirely different message from the original speech that this was lifted from. As fake news becomes more of a concern, the prospect of creating a highly convincing yet entirely fabricated speech from a powerful leader seems disturbingly easy and could lead us to a dark reality that isn’t far off of Charlie Brooker’s ‘Black Mirror’episodes.
We often hear that the robots will take our jobs with the intention of the greater good, alleviating us from mundane tasks and freeing us up to be more creative and to develop our own skills, whilst leaving those repetitive tasks to the machines.
We experienced a number of these AI powered technologies in action at the exhibition, from AI powered hovers to a robotic armed bartender. We tried a cocktail from our synthetic friend and I can tell you, it was strong! Although the flailing arms of the mechanical Lloyd was amusing to watch, it lacked the traditional qualities and charm you would expect where human interaction is a part of the experience. We’re used to asking the barman to serve our drink to our own particular taste, receiving a friendly smile and pleasantries, bending their ear and building a real human connection, which lacked from our robotic mixologist.
AI is already helping us in fantastic ways through health care, finance, industrial solutions, however I think within social situations, where the true experience comes from the people you interact with. A robot, no matter how lifelike, will not replace the human connections we crave in society and social situations anytime soon, but it was a fun experience and I’m sure ‘I’ll be back’.