AI UK 2022: conversation, innovation and inspiration

We look back at the highlights of the Turing’s largest ever event

Thursday 21 Apr 2022

On 22-23 March, some of the leading figures in artificial intelligence gathered at the Park Plaza Hotel in Victoria, London for AI UK 2022, the UK’s annual showcase of data science and AI.

This year’s hybrid event was broadcast live around the world, and this was a truly global showcase, with more than 1700 attendees joining from 53 countries. It was our largest ever event, with our most diverse and expansive programme to date: over 200 speakers delivered 50 sessions and 11 unique ‘Research in action’ interactive workshops.

diversity book club

This year's hybrid event featured some speakers live on set, with others joining virtually from around the world along with attendees. Here Anjana Ahuja (pictured left) is joined by Steph Wright (pictured right) for the Diversity in AI book club session

The sheer size of the programme meant that a huge number of topics were tackled by our speakers, with certain themes emerging across the four stages. While AI UK 2022 was emphatically a celebration of all the brilliant work being done in the field of AI and data science, there were also notes of uncertainty and urgency sounded by our speakers, reflecting the scale of the challenges faced by humanity, from conflict and climate change to disease and widening inequality.

Several speakers stressed the need to get away from notions of AI as necessarily threatening, or on the other hand as inherently utopian: a “magic bullet” to counter all future challenges, as National Statistician Sir Ian Diamond put it. Technology is not inherently good or bad, the author Jeanette Winterson argued, but it is up to us to take responsibility “for the direction of travel”, as she cautioned against a world in which AI is increasingly harnessed for labour exploitation, weaponry, and surveillance. She drew an analogy with her hometown of Manchester– once dubbed “the golden sewer” for the way in which vast wealth flowed from the misery and poverty of workers in the 19th century– and warned that AI has the potential to produce such extreme disparities again “if we’re not careful”. And using a strikingly similar metaphor, the Oxford Internet Institute's Callum Cant noted that factories full of badly-paid “ghost workers” performing menial tasks for big tech firms are the “dark satanic mills” of our age.

Hosting a lively panel asking whether AI has actually delivered on its promise to society, The Economist's Kenneth Cukier asked whether those who misuse technology might provide an impetus for “the people who are going to counterattack”, acting as “fuel” for further innovation. A great example of this innovation was presented on the Public policy stage, where a team of Turing researchers showcased the Online Harms Observatory, an AI tool to help fight online hate speech, developed in collaboration with the Department for Digital, Culture, Media & Sport.

As one of the project’s developers, the Turing's Head of Online Safety Bertie Vidgen noted, the Institute is very well suited to delivering projects such as the Online Harms Observatory, which combine “the rigour and trustworthiness of academia” with the “speed and flexibility” of being unconstrained by academic publishing cycles, instead collaborating directly with the government and civil society groups for an immediate, real-world impact. The Public policy stage also featured a four-part session on international collaboration in AI and data governance organised with the Global Partnership for Artificial Intelligence (GPAI), who were one of AI UK's main event partners.

Speaking on the Public policy stage, Sir Patrick Vallance, the government’s Chief Scientific Adviser, commented that during the COVID-19 pandemic, we became “a nation of armchair epidemiologists”, anxiously interpreting the latest statistics on infections from our living rooms, and becoming ever more fluent in concepts such as exponential growth and the R value. Sir Patrick noted that we have learned a great deal from this pandemic, and that AI programmes like DeepMind’s AlphaFold have the potential to develop vaccines, diagnostics, and therapeutics within 100 days of the next pandemic hitting, though he cautioned that the next “existential crisis” may not come from pathogens.

Over on the Climate action stage, minds were focused on how AI might help us mitigate a different threat to humanity. One example came from Emma Bland, Professor of Environment and Human Health, who demonstrated an ingenious ‘Local Climate Adaption Tool’ (LCAT) which can be used to zoom in on specific geographical regions and, using local climate data, generate precise predictions about climate impact on everything from transport disruption to air and water pollution.

Such demonstrations served as a reminder that AI can be used as a practical tool, and is not just a futurist fantasy – a point that several of our speakers were keen to emphasise. Tania Duarte from the Better Images of AI project pointed out how “alienating the sci-fi image of AI” can be, and the need to make AI “culturally and contextually relevant” to our present moment. This is something that Better Images of AI has done with its library of images, which aims to distance AI from the robots-and-glowing-brains clichés.

Across the four stages, speakers from different disciplines discussed the extraordinary potential of machine learning, not just to fight existential threats, but also to help humans to flourish and better understand our world. As Sana Khareghani, Head of the UK Office for AI put it, "it really is magical the kind of things these technologies can do." On the ‘Reimagining the past with AI’ panel, Turing Research Associate Mariona Coll Ardanuy argued that algorithms should be used “not to simplify reality”, but “to enrich it”. On the same panel, Senior Lecturer in Digital History Melodee Wood looked forward to a future where hundreds of different algorithms can be used to interpret a single subject, just as there are hundreds of different single-topic books today.

The last word should go to Jeanette Winterson, whose session proved among the most popular at this year’s event. As we make ever greater advances in technology and science, we should not lose sight of our humanity. Our “prodigious intelligence” should be married to a sense of love, “heart and head together”.