PD Newsletter #85: Celebrating UK's female athletes

👋🏽👋🏻👋🏾 Hello, welcome.

This month we are thrilled to share our sponsorship of Hackney District Girls football team as part of our long-term partnership with Sporting Hackney FC. It’s a pleasure to be able to promote and celebrate the emergence of girls’ football, and we wish the team the best of luck in their upcoming matches.

We’re also sponsoring our friend the fearless Martha Lane Fox on her mission to climb the UK’s three highest mountains in support of four charities.

Rosemary and Emma

🚀 Forwarded this email? Sign-up. Follow us on LinkedIn

Latest from Public Digital

Watch our last PD Session on the person-centred health record. And read a summary of our Session earlier this year on digital transformation in federalised governments.

Last month we were visited by Andrew Dudfield, Head of AI at fact checking organisation Full Fact, who spoke to us about using AI to counter ‘bad’ information.

We’ve been celebrating our PD Network with two mixer events (pictured below), giving us the opportunity to honour the hard work of our amazing freelance consultants. With their help, we can assemble the best possible teams to tackle complex challenges. Listen to Emma Parnell talk about being in the PD Network.

Ways of working

👏 Either/Org are working on a practical framework for organisational design, the first of its kind to be developed by women and people of colour. For more taxonomies, this fantastic team taxonomy by our Network member Emily Webber breaks down four different classifications of what makes a team.

🚫 The cost of failure in AI: Tom Loosemore writes that where we might be impressed by the 80% success rate of an AI chatbot, the implications of the 20% failure rate are what really matter. “How many people realise the answer is wrong?... At what cost to them and to you?” Sarah Gold responds with a challenge to transform AI’s unreliability into a feature, not a bug, by creating design patterns for AI which include a “show confidence” pattern.

🔍 What to do with user research findings? This piece on using guiding principles by Vicki Riley at Coop illustrates the power of translating core findings into guiding principles. These principles work because they are actionable, memorable and bring the team together around a shared problem.

🔬 Lots of good stuff in breaking down hypotheses to make them easier to test by Jeff Gothelf. It serves as a reminder that testing hypotheses mostly comes down to testing the value of the idea for users: “If no one wants it, it doesn’t matter how hard or easy it is to build it.”

🇮🇸 Content design gone wrong by Anna Andersen: How you can accidentally run for president of Iceland. This is thanks to the country’s new digital candidate endorsement system whose webpage confused users into mistakenly registering as a candidate. The lesson for UX designers is simple: “don’t make people think or read too much”.

🏳️‍🌈 Finally, read this deep dive into designing pronoun selectors by Chelsea Roden for some useful tips, like asking for ‘pronouns’ rather than ‘preferred pronouns’, and why you should steer clear of the option ‘other’.

State of technology

💀 The rise of Chinese AI ‘deathbots’: These AI avatars of deceased loved ones, which use LLMs as well as dystopian-sounding cloned voices and appearances, are reportedly helping people grieve. Their popularity in China is attributed to the country’s strict controls over religion and spirituality, which have led to a lack of “publicly available resources for bereavement”.

🗳️ AI is being used to resurrect the dead for different reasons in India, where some election campaigners are using ‘soft fakes’: AI generated videos of deceased political figures, to influence voters. For more on AI and elections, Rest of World’s elections AI tracker reports on incidents of AI-generated election content globally. Also check out how AI detection tools work (and how they don’t work), and a tool for identifying which chatbots are trustworthy.

🌟 Heartening news that Taiwan managed to successfully counter disinformation during its January elections. TLDR: media trust among the population matters, earned through accurate, short-form reporting and direct debunking of disinformation. Mexico is tackling the same problem in the lead up to their elections with an AI chatbot which citizens can access via Whatsapp to query and verify news.

💔 Ed Zitron on how Google search died: the story of Google’s decision to lower the quality of its search function for the sake of growth-at-all-costs, making it “less reliable, less transparent, and dominated by search engine optimized aggregators, advertising, and outright spam”. If the doom-and-gloom leaves you nostalgic for the early web, read Molly White on how we can get it back.

UK election watch

⏳ The UK’s technology trade association published its Seven Tech Priorities for the next government in relation to digital UK.

🔄 Paul Maltby weighs in on the ever growing commentary about a local GDS. Having established and run the Department for Levelling Up’s local digital team, he offers some learnings for digital service reform from a local government perspective, including what that reform should - and shouldn’t - focus on.

🚀 Rob Hopkins gives us a taste of policymaking underpinned by radical imagination in the Ministry of Imagination manifesto. Adapted from the 100 episodes of the ‘From What If to What Next’ podcast, it offers an imagined future with radical reform in everything from government to employment to lifestyle. Food for thought, at the very least.

Digital government

🇳🇬 Nigeria’s government has announced the launch of its first multilingual LLM, a notable departure from a landscape of English-language dominated LLMs. They’ve addressed the complication of Nigeria’s few hundred different native languages by choosing to train it on five low-resource languages and English in local accents. The government also confirmed that it has secured $2 million in funding to begin acquiring GPUs for various AI projects.

🔥 You have to take risks to find value: In this webinar on how the UK can encourage the uptake of AI in the public sector, Dave Rogers argues that the need for agile, test-and-learn approaches in the public sector becomes most acute when it comes to AI (12:02). Government should be an innovator in AI, but strict procedures in public spending make experimentation challenging.

💰 This new podcast examines the Phoenix payroll disaster, a failed government payroll transformation which affected thousands of public servants in Canada. The story bears the hallmarks of other failed transformations: a failure to do the hard work early or design for the complexity of the service. “Some people thought payroll was simple. And it just is not.”

💍 Ukraine has taken huge strides in digital government, including introducing an AI generated foreign ministry spokeswoman or ‘digital person’, modelled on a real Ukrainian singer. They are also trialling online marriage via the Ukrainian Diia app, aimed at couples separated by the ongoing conflict.

🎓 The Biden administration is following in the footsteps of the UK government in scrapping degree requirements and focusing on skills-based hires for its tech workforce. The aim is to open up work to more job-seekers and give the government greater access to tech and AI talent in a competitive market.

🤝 And finally, the UK parliament has launched its new Information and Digital strategy. Bringing information, data and digital into one strategy for the first time, it is underpinned by the principles of collaboration, iteration and openness.

Something fun

🐵 This UX timeline tracks the design evolution of brands from Spotify to Hubspot, including the dramatic upsizing and downsizing of the Mailchimp monkey.