People's Panel on AI Bulletin 5 - Recommendations and Review

This is the fifth People’s Panel on AI Bulletin, and our last daily update - covering the Panel’s presentation event. You can find earlier bulletins from the week here. Look out for one more update soon with the final report, and please do take a few minutes to provide your feedback through the anonymous evaluation survey.  

Recommendations

After a week of deliberation, engaging with experts, discussing lived experience, attending events at the AI Fringe, and following updates from the UK AI Safety Summit, the members of the People’s Panel on AI, drawn from across England, recommend:

• (1) A global governing body for AI.

  This should bring together citizens, impartial experts and governments from across the world, and to ensure regulatory collaboration that includes the global south.

• (2) A system of governance for AI in the UK that places citizens at the heart of decision making.

  Roundtables of scientists, researchers, ethicists, civil society, academia and industry should inform and provide evidence for government and citizens to then work with roundtables on decisions.

• (3) Awareness raising about AI across society.

  From the classroom to the home. From the workplace to the community. This should highlight risks such as addiction to social media, as well as the opportunities that AI offers.

• (4) A safe transition, with training, to support people into a world of work alongside AI, ensuring no-one is left behind.

  This could include a tax pot dedicated to training and reskilling, funded by employer contributions.

• (5) A continued national conversation on AI, including retaining the People’s Panel to keep public voice live in a fast-changing AI landscape.

  We citizens can do jury service and as such are already trusted to make life-impacting and significant decisions.

• (6) Focus on inclusive collaboration, to set out a vision of life where AI is used to enhance and balance human needs.

• (7) Stakeholders acting with transparency at all times.

  An example of this might include a ‘black box flight recorder’ approach to AI models, which protects intellectual property, but can be shared openly when things go wrong. 

You can find a recording of the panel presenting these recommendations available to watch here, and a copy of the slides can be found here.

Please do spare 15 minutes to watch the Panel’s presentation, and then you can also catch on the recording responses to the recommendations from attendees from TechUK, Mozilla, Ada Lovelace Institute, Google and others

“This initiative, this People’s Panel, has been incredibly important. There have been a ton of AI events that I wanted to get to this week, but this was the most important one for me to come to, because I wanted to hear what you had to say. 

In terms of your recommendations, first of all, well done. A list of seven recommendations in this field is really impressive. You really got down to the most important points. Some of those points I think are very much part of the debate, but I think the crucial point that you’ve made time and again is the need to involve people in these discussions. 

AI is about technology and it’s about people. We hear a lot about technology and we hear a lot from the technologists and the business leaders and the politicians and the academics, but we haven’t heard enough from the people.” Antony Walker, Deputy CEO, TechUK

And as lead facilitator Henrietta Hopkins, of Hopkins van Mil, explained, these recommendations are just the tip of the iceberg of panel discussions. In the coming days the whole team will be collating notes from the last four days, and producing an additional report of the discussions, which will be in your inboxes soon.

Review

As part of their deliberation, this morning the Panel also reflected on the strengths and weaknesses of the AI Fringe and Safety Summit in representing the concerns of members of the public.

We were glad to hear…

• A focus on creative industries, education, jobs and elections

• That the Bletchley Declaration acknowledge importance of AI safety - and analogies to other safety systems (e.g. nuclear and airlines)

• The need for regulation and emphasis on governance

• The politics of AI, and that not everyone thinks this is a good idea

We wanted to hear more…

• Meaningful focus on AI and climate

• Focus on the power that technologists hold, and who should have power

• Current harms of AI, particularly through social media

• Practical safety systems to address AI as a possible weapon of mass destruction

• How citizens should be involved, and how we can democracy at all levels

As the UK passes the baton to South Korea and France for future AI Safety Summits, can the public have a greater role in setting the agenda: taking safety seriously both in the here-and-now, and for the long-term?

Over to you…

At the end of today’s presentation event, I asked everyone in the room and online four questions: 

• What have you heard that resonates with you?

• What implications do you think these recommendations have?

• Are there any messages or recommendations from the People’s Panel you can commit to championing?

• How could you embed public voice in your work in future?

If you have a response to add please drop a message to tim@connectedbydata.org with your thoughts. 

I will be collating responses over the next week to share with the members of the People’s Panel on AI, and we’re committed to helping the voices raised through the Panel be heard as widely as possible. 

Thank you

I want to close this bulletin with a note of thanks to all the members of the panel who have worked so hard this week: Margaret, Joe, Ollie, June, Sharif, Janet, Elizabeth, Shanti, Ermias, Sallie and Adam. 

The Panel, and all the team who have supported them this week, are heading for a well-deserved weekend. But we’ll be back in touch soon with the final report & news on possible next steps.

(And if you’ve read this far, that evaluation survey link once again: https://www.smartsurvey.co.uk/s/D5XKBH/ - we really would value your feedback!)