Hello ODI Supporter,
London’s Metropolitan Police announced this week that it will expand its facial recognition software into the West End by Christmas, followed by a further six areas. The Met trialled the technology in Croydon, where its use led to 173 arrests, including a woman who had been wanted for 20 years. Alongside the arrests, the force claims it saw a crime reduction and a significant fall in violence against women and girls. Over 470,000 faces were scanned in the trial, with apparently only one person misidentified. But opponents of the scheme say that it forces people to enter a digital police lineup, while the algorithm used to power the system can discriminate against Black people. According to the Met, typical uses of the software include real-time location of people on a ‘watchlist’, identifying people who are injured or unconscious, and retrospectively identifying people with images held on databases.
In a week that saw China top the list for having the most powerful computer in the world, the Chinese e-commerce and tech platform Alibaba has been accused by Anthropic of extracting capabilities from its Claude AI model. In a letter to two members of the US Congress, the tech giant accused Alibaba of being behind nearly 29 million interactions with Claude using thousands of fraudulent accounts to “illicitly extract Claude's capabilities". Known as distillation attacks, answers are extracted from a stronger AI model to train a weaker one, in this instance targeting Claude’s ability to tackle longer and more complex tasks and its approach to decision making. OpenAI has previously accused Chinese firms of doing the same.
This week, we published a new report to help businesses gain the benefits of using AI by helping them to get their data foundations AI-ready. This is our first piece of research for the Interchange for Data and Enterprise AI (IDEA), sponsored by SAP and hosted by the ODI. IDEA is a vendor-neutral, community-led programme that advances the global conversation on enterprise data architecture and AI readiness. IDEA will set up a programme of research, community building and peer learning to help enterprises of all sizes make the most of the promise AI holds through AI-ready data infrastructure and well-governed enterprise data. Learn more and find out how you can get involved.
Don’t forget you can still get tickets to hear our Executive Chair Prof Sir Nigel Shadbolt cut through some of the critical issues in AI on Thursday 16 July, 16:00-17:00. And on Thursday 30 July 12:00-13:00, we have Rowenna Fielding, privacy, data protection and data ethics geek for hire (her words, not mine!), as she discusses her data framework for improved data governance. Get your tickets now!
And finally… It’s hot. It’s very hot. Have you noticed that? Crikey, it’s warm. And if you’re looking for some reading material for when you’ve shut all your curtains and windows and you want to take your mind off the room temperature, you can explore more than 1,600 sets of meteorological, magnetic and tidal observations taken from 1706-1915 from across the globe that have been digitised for the first time by the Royal Society. Will reading historic weather data keep you cool? It will do with an ice lolly.
David and Jo
PS: Our Senior Researcher, Calum Inverarity, wrote a piece for the Chatham House publication The World Today looking at the threat posed to internet infrastructure by recent conflicts. Check it out.
