Exploring our relationship with AI: highlights from the Barbican exhibition
Last week, some of our tech team caught the ‘AI:More than Human’ exhibition at the Barbican, London. From human interaction with inanimate objects, through to Babbage, Lovelace, Turing, and current developments in agriculture, healthcare, social mobility and robotics, the exhibition explored our relationship with Artificial Intelligence over the ages and the blurring of boundaries between technology and human life.
Entering the exhibition space to rather grating, chilling background ‘music’ (imagine nails down a blackboard!) created a sense of unease – was AI going to take over humanity!? Hundreds of different interactive AI scenarios led to a sensory overload and a slight feeling of chaos.
But overall, the narrative it provided around AI was positive and thought provoking. The exhibition included some fantastic tangible examples of where AI is truly benefitting our health and society, but certainly did not ignore the ethical questions around the wider adoption of AI.
With the exhibition now over, we thought we’d share some of the things that stood out to us the most:
1. We met the world’s best conversational chatbot
Most of us are already interacting with chatbots every week (sometimes without being aware) so we were interested in putting ‘Mitsukuto’ the test. Created from AIML technology by Steve Worswick, a senior AI designer for chatbot developer Pandorabots, this chatbot has won four awards at the Leobner Prize Turing Test and is positioned as a virtual ‘friend’ rather than an assistant.
Could this be one step to addressing loneliness? For now, we’re not convinced.
Conversation felt quite forced and we had to keep it simple – one sentence at a time. But her AI is allowing her to learn from every conversation she has (and you can directly teach her when she makes mistakes), so she’ll only get smarter. You can test her out – and teach her – here (Flash needed).
2. The issue of ethics in AI
There were many displays which had an undercurrent of the issue of ethics in AI, but one which particularly stood out was the work of MIT researcher Joy Buolamwini. Through her research she has found that many AI developments had a gender bias towards white males. Many facial recognition AI technologies are not able to classify women, and black women in particular, correctly. Her display used examples of famous black women to demonstrate the technology’s failings.
Of course the challenge is that ultimately, AI stems back to data which is selected and prepared by humans, and our bias – conscious or otherwise – is being translated into AI algorithms for hundreds of applications. With the ethics of AI becoming increasingly under focus, we expect a lot of leading data scientists and AI organisations to urgently look at ways of addressing this.
3. AlphaGo outmaneuvers a human brain
Many of you will remember when Google DeepMind set the challenge for its computer software AlphaGo to beat South Korean Go grandmaster Lee Sedol – and this achievement was one of the central video displays at the exhibition. Go is a game of strategy and intuition, with endless tactical possibilities within the game, which is why it is said to be much harder for an AI to beat its human opposition than in a game of chess. This was of such significant importance within AI development because:
- Ultimately Lee lost 4-1 to an AI which was the first time this had happened to such a seasoned player
- It was a huge leap in the development of AI in a game known for intuition; which isn’t a traditional computing strength
- It redefined the game, pushing Lee Sedol way beyond his comfort zone and together they created new strategies and moves that had never before been seen in the game’s history
4. AI and creativity
Can AI be truly creative? This question has troubled computer scientists, artists and philosophers for over a decade. Using its ability to learn human creativity from vast datasets, AI has the potential to disrupt our notion of creativity.
The exhibition offered a unique glimpse into the power of AI to create meaningful stories. Excerpts from poems and newspaper articles asked us to decide whether they were written by AI or by a human. Interestingly, most of the content that seemed to be created by a human was actually written by AIs.
This extended into a look at the future of journalism. Will our future reporters and writers be AIs? While financial reports and media brands such as Forbes and Washington Post are already using AI to create some stories, only time can tell whether AI will replace human writing or simply supplement and augment it.
5. AI and the natural world
AI is typically seen as the domain of technology and robotics, not as a tool for shaping the natural world. But the exhibition surprised us with a few exhibits that made us question this notion. While AI emulates the behaviour of the brain, the related research area of artificial life (A-life) explores a much wider set of natural processes, including human and animal biology and environmental science.
The exhibition included a display of a controlled agricultural environments where AI is used to create and maintain the optimum conditions required to grow pretty much any kind of crop. This artificial farming process could be useful for addressing foot shortage and supply issues arising from the damaging impact of global warming and overpopulation.
This exhibition was an excellent way of educating – and opening up people’s minds to – the possibilities of AI. We left the exhibition with far more questions than we came in with. Who should be responsible for setting ethical standards with in AI? What influence will we – as the general public – have in setting these standards? What happens if (or when) AI gets into the wrong hands?
But it is by asking these questions, and starting to explore the answers in such a public setting, that more people will be able to understand, accept, and hopefully embrace, the benefits AI can bring to the economy, society, health, the environment and beyond.
*Image credit: AI: More than Human, Barbican Centre – 16 May-26 August 2019. DeepDream: The Artificial Pareidolia © Alexander Mordvintsev Credit Tristan Fewings/Getty Images
This blog was written by the Brands2Life Business Technology team