Helen Margetts

@helenmargetts.bsky.social

Professor at the Oxford Internet Institute, University of Oxford and the Alan Turing Institute, trying to make sense of governance and democracy in a digital world #public policy #digital government #data science #AI #democracy #behavioural science

The notion that child sexual abuse imagery is on a par with “woke” political criticism is patently absurd. Regulating sexual content online "will be a game of catch-up, like every technology spiral before, but it will have to be played." By @helenmargetts.bsky.social

Regulating Sexual Content Online has Always Been a Challenge. How we got Here

It takes something really horrible for policymakers to take swift action.

politicalquarterly.org.uk

The notion that child sexual abuse imagery is on a par with “woke” political criticism is absurd. Regulating sexual content online "will be a game of catch-up, like every technology spiral before, but it will have to be played." @helenmargetts.bsky.social politicalquarterly.org.uk/blog/regulat...

Regulating Sexual Content Online has Always Been a Challenge. How we got Here

It takes something really horrible for policymakers to take swift action.

politicalquarterly.org.uk

"The risk of inequity is always there, wherever you've got AI" Anya Pearson asks @helenmargetts.bsky.social, Professor of Society and the Internet at the Oxford Internet Institute, about digital government and the opportunities and challenges of AI.

"The Risk of Inequity is Always There, Wherever You've Got AI": Interview with Helen Margetts

Anya Pearson interviews Helen Margetts on the relationship between digital technology and government, politics and public policy.

politicalquarterly.org.uk

New paper alert! ‘The levers of political persuasion with conversational AI’, lead author @kobihackenburg.bsky.social co-author @helenmargetts.bsky.social @oii.ox.ac.uk @socsci.ox.ac.uk @ox.ac.uk.

Kobi Hackenburg@kobihackenburg.bsky.social · last yr.

Today (w/ @ox.ac.uk @stanford @MIT @LSE) we’re sharing the results of the largest AI persuasion experiments to date: 76k participants, 19  LLMs, 707 political issues. We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!  🧵:

I am so proud of the brilliant @oii.ox.ac.uk DPhil @kobihackenburg.bsky.social and this wonderful “bees-knees” paper on conversational AI and political persuasion @AISecurityInst - it is a “must-read”, comments welcome!

Kobi Hackenburg@kobihackenburg.bsky.social · last yr.

Today (w/ @ox.ac.uk @stanford @MIT @LSE) we’re sharing the results of the largest AI persuasion experiments to date: 76k participants, 19  LLMs, 707 political issues. We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more!  🧵:

Those big, beautiful eyes! A tiny owl braving the winter snow, perfectly camouflaged in its snowy perch. Nature's little marvel. 📸 mon_regard_sur_la_nature

A Northern Saw-whet Owl, perched on a snow-covered branch in a close-up view. The owl has large, bright yellow eyes and mottled brown and white feathers, blending into the snowy background.

I just want to tip my hat to the crack team of White House economists who were able to discover--in just a few short days--that the U.S. is dependent on China for smartphones, computers and semiconductors.

This morning my phone asked me insistently whether I wanted to send a ‘running late’ message to the person I am meeting at lunch time tomorrow. Is that predictive capability - will I be late? Or is it a sign that the age of useful agentification is not quite here?

A lovely two days @wepsocial.bsky.social Anniversary Workshop, thank you everyone!

West European Politics@wepsocial.bsky.social · last yr.

We are very happy to host the WEP 5️⃣0️⃣th Anniversary Workshop @lmumuenchen.bsky.social today & tomorrow! Great presentations, intense debates & we are grateful for colleagues all around the globe joining us. Sessions on: 👉elections, parties, mobilization 👉welfare politics 👉🇪🇺 👉policy & administration

Picture of the participants of the workshop.

📈Out today in @PNASNews!📈 In a large pre-registered experiment (n=25,982), we find evidence that scaling the size of LLMs yields sharply diminishing persuasive returns for static political messages.  🧵:

BildBild