Gaia Marcus

@gaiamarcus

Director @adalovelaceinst.bsky.social:ensuring data & AI work for ppl & society Stint in government - led #NationalDataStrategy; roles in Cabinet Office, ONS & MHCLG Charity roles inc. Samaritans Trustee; staff @ The RSA, Centrepoint, ParkinsonsUK

2.722 Folgt11.226 Follower254 Posts

Website Bluesky

Ist das dein Profil?Übernimm es — dann gehört dir gaiamarcus.dirty.social.Profil übernehmen

Gaia Marcus@gaiamarcus· Mar 18

Very thoroughly earned vallor (!) - @shannonvallor.bsky.social given the Ai & Robotics Research Leadership Award!

Gaia Marcus@gaiamarcus· Mar 18

Fabulous to hear from @adalovelaceinst.bsky.social board member @shannonvallor.bsky.social at tonight's Responsible AI Robotics and AI awards - I'm always so inspired by her morally crisp and targeted calls for AI that centres people, augmenting our agency and thinking, rather than supplanting it

Shannon Vallor's talk - ' The machine starts: taking stock of the AI and Robotics revolution '

It's a play on E.M. Forster's The Machine Stops
-
“THE MACHINE IS MUCH, BUT NOT EVERYTHING."

Shannon's call: WE NEED NEW AND BETTER AI/R STORIES:

Stories about Al and robots beyond GenAl chatbots and agents

Stories about humans using their own brains more with Al

Stories about Al and robots that tread lightly on our lives and minds

Stories about Al and robots that help mend social ties & trust

Stories about Al and robots that repair parts of our world & institutions

Stories about Al/R researchers and builders who ask and listen

Stories about people enabled to choose the kind of Al/R they want

Gaia Marcus@gaiamarcus· Jan 14

📢 This isn’t an unpopular idea: nearly 9 in 10 people in the UK want independent AI regulation. Yet the current oversight of AI falls far behind that of other sectors (like aviation, pharmaceuticals and financial services), with no clear plans for improvement.

Stat picture card - 89% of the UK public say it is important to regulate AI independently

Gaia Marcus@gaiamarcus· Jan 14

⚖️ We cannot stay ahead of these harms without robust AI regulation. Currently, we are chasing after individual tragedies and scandals, attempting to plug the gaps with existing laws and regulation (image). This isn’t enough: we need to manage harms at the source, and not just manage the symptoms.

Ada graph of sector regulation which shows how absent AI regulation is, compared to a range of other sectors

Compared to the following:

Aviation

Financial Services

Pharma

Food Safety

Foundation models/general-purpose Al lack regulation lack real coverage beyond voluntary standards across:

Proactive risk monitoring

Safety standards

Independent standards
Market entry authorisation

Post-market monitoring

Independent regulator

Enforcement powers

Accountability measures

Transparency/reporting requirements
Routes for redress

Gaia Marcus@gaiamarcus· Dec 4

Social card - 89%

of of the UK public believe Al products and services should not be rolled out until they are proven safe, even if this slows things down

Source: A nationally representative online poll with 1,928 UK adults aged 18+ between 8-24 September 2025

91% of the public think it is important that Al systems are developed and used in ways that treat people fairly

Source: A nationally representative online poll with 1,928 UK adults aged 18+ between 18-24 September 2025

89%

of the UK public believe government and regulators should have the authority to halt the use or sale of any Al system found to cause harm

Souurce: A nationally representative online poll with 1,928 UK adults aged 18+ between 8-24 September 2025

84% of the UK public are concerned about the government putting technology sector needs ahead of the public when regulating Al

Source: A nationally representative online poll with 1,928 UK adults aged 18+ between 8-24 September 2025

Gaia Marcus@gaiamarcus· Oct 22

A full day of AI and [insert topic] panels, lunch, conference and dinners today, across London and Cambridge. Will I need the sign?! /Via @xkcd.com / @wikipedia.org

Protester holding up 'citation needed' sign, likely at a protest.

From https://es.wikipedia.org/wiki/Archivo:%22Citation_needed%22.jpg

Gaia Marcus@gaiamarcus· Oct 20

Will AI solve our woes? At SciTech Committee I highlighted - 🍏 or 🍊? Need specificity when we say 'AI' - 🔬 narrow/specific AI usecases used to justify invsment in Gen AI - 👩‍🔧 Dnt forget the plumbing! AI pilots =\= solving legacy IT & data issues - 🏛️ AI What Work Centres cld help learn real lessons

Gaia Marcus@gaiamarcus· Oct 20

Delighted to give evidence to @chionwurah.bsky.social and the UK parliament Science and Technology Committee last week I argued that "the public hold the public sector to a higher account because they do not have the ability to opt out" You can't just 'shop around' when accessing vital services

Gaia Marcus@gaiamarcus· Sep 24

KD rallying the troops for what will happen next - can't wait to see what these three amazing organisers do next!

Gaia Marcus@gaiamarcus· Sep 24

Amazing panel with a cross-section of the people doing grounded, community-centered organising KeShaun Pearson - Memphis Community Against Pollution Marisol - No Desert Data Center KD Minor- Alliance for Affordable Energy Amba Kak @ainowinstitute.bsky.social

On the panel:
KeShaun Pearson (Memphis Community Against Pollution) will break down how Elon Musk’s xAI supercomputer is polluting the air of historically Black neighborhoods in Memphis, and how organizers are fighting back against yet another chapter of corporate extraction in their communities

Marisol (No Desert Data Center) will share about their grassroots coalition’s recent win in Tucson to stop Amazon’s Project Blue data center proposal, which threatened the city’s scarce water supply, and how they’re organizing for future protections.

KD Minor (Alliance for Affordable Energy) will demystify the energy impacts of data centers in Louisiana and share organizing strategies to mobilize community opposition to Big Tech and Big Oil infrastructure.

Amba Kak (AI Now Institute) will talk us through the bigger picture: what’s behind Big Tech’s AI data center expansion, who stands to benefit from this boom, and what we sacrifice in return.

Gaia Marcus@gaiamarcus· Sep 24

Over in Brooklyn for a live @themaybe.org podcast - Let Them Eat Compute: Computer Says Maybe Climate Week Livestream - @ambakak.bsky.social and @alixdunn.com setting the scene for a conversation 'rooted in the material, in the current context, in activism'..

Gaia Marcus@gaiamarcus· Sep 17

I know not everyone loves a pdf (html and accessible formats are also available), but she is *pretty*

Strategy excerpt - what we have learnt Our strategy 2025-28

Data is increasingly used, collected and processed in ways that were previously unimaginable, for example, biometrios capabilities which evolved from verification, identification and categorisation systems to emerging systems on cognitive and biometric inference. This means that regulatory and governance regimes can fall out of step with practice on the ground.

In many cases, there is not enough evidence to know whether or in which contexts Al and data-driven systems work as Intended. This is compounded by low levels of transparency and a lack of publicly available evaluations. Society and public services are often over-reliant on industry accounts of innovation and opportunity.

People hold diverse and nuanced views about Al that are not always taken into account by governments, companies and other powerful actors. The tools that have historically emerged to allow people and communities to steer innovation-regulatory checks and balances, civil society and trade union mobilisation, and political leadership from elected representatives - are not always working well enough to protect people and society from the negative impacts of rapid technological change. New participatory mechanisms and deliberative approaches hold promise but require institutional backing and political buy-in to succeed.

What we have learned

Over the last seven years, Ada has established itself as a trusted, evidence-based institution.

Independently funded and curious, we have sought diverse perspectives on data and Al and successfully brought together evidence and conversations from across civil society, academia, governments, the technology sector and the public. Our measured and balanced voice provides clarity and rigour in a landscape often dominated by hype, hope or fear.

Our work exploring the intersection between Al, data, people and society has surfaced the following lessons:

Al and data-driven technologies are 'sociotechn…

Our strategy 2025-28

Data is increasingly used, collected and processed in ways that were previously unimaginable, for example, biometrios capabilities which evolved from verification, identification and categorisation systems to emerging systems on cognitive and biometric inference. This means that regulatory and governance regimes can fall out of step with practice on the ground.

In many cases, there is not enough evidence to know whether or in which contexts Al and data-driven systems work as Intended. This is compounded by low levels of transparency and a lack of publicly available evaluations. Society and public services are often over-reliant on industry accounts of innovation and opportunity.

People hold diverse and nuanced views about Al that are not always taken into account by governments, companies and other powerful actors. The tools that have historically emerged to allow people and communities to steer innovation-regulatory checks and balances, civil society and trade union mobilisation, and political leadership from elected representatives - are not always working well enough to protect people and society from the negative impacts of rapid technological change. New participatory mechanisms and deliberative approaches hold promise but require institutional backing and political buy-in to succeed.

What we have learned

Over the last seven years, Ada has established itself as a trusted, evidence-based institution.

Independently funded and curious, we have sought diverse perspectives on data and Al and successfully brought together evidence and conversations from across civil society, academia, governments, the technology sector and the public. Our measured and balanced voice provides clarity and rigour in a landscape often dominated by hype, hope or fear.

Our work exploring the intersection between Al, data, people and society has surfaced the following lessons:

Al and data-driven technologies are 'sociotechnical". This means that technologies do …

Gaia Marcus@gaiamarcus· Jun 12

. @mbirtwistle.bsky.social Does the UK need an AI Act? AI already embedded in how we work, communicate & access public services, but still no coherent legal framework to govern use Instead, we rely on fragmented patchwork of rules, while AI developers mark their own homework 👉 tinyurl.com/yc883jcc

Gaia Marcus@gaiamarcus· Jun 12

Me in @newstatesman.com on AI use in the public sector. 🔹Public trust is essential 🔹 AI won’t magically fix entrenched problems 🔹 Supply chain, lack of evidence & risk of uneven capabilities need attention 🔹 72% of UK public say laws & regulations wld increase comfort with AI 📖 tinyurl.com/2ut88e48

An image of the print version of https://tinyurl.com/2ut88e48

Gaia Marcus@gaiamarcus· Mar 26

Now hearing from Turning/ @oii.ox.ac.uk's @helenmargetts.bsky.social on the @adalovelaceinst.bsky.social /Turing survey on uk public attitudes to AI attitudestoai.uk Public have nuanced views, expectations of regulation up to 72% and LLM awarenesses abd use ↗️rapidly (note small % of habitual users)

About the survey

• Wave 2 of a tracker survey: How do people feel about Al?

• Sample of 3,513 UK participants using random probability sampling

• With overrepresentation of Black/Black British, Asian/Asian British, low income and low digital skills participants

• Attitudes towards 8 different applications of Al, & governance and regulation of Al

Facial recognition for policing

Assessing eligibility for welfare benefits

• Assessing risk of developing cancer from a scan

Assessing risk of repaying a loan

Driverless cars

Robotic care assistants

• Large language models

Mental health chatbots

People see benefits in Al, but concern about some applications of Al is increasing

Attitudes vary by application

86% see benefit of Al to detect cancer risk

75% concerned about driverless cars

People can see both the benefits and risks

Even though majority see benefits of Al in cancer detection, 64% still worried about relying too heavily on technology rather than professional judgements

Concern is rising

59% are concerned about the use of Al to determine eligibility for welfare benefits such as Universal Credit. This figure was 44% in 2022/23People see benefits in Al, but concern about some applications of Al is increasing

Attitudes vary by application

86% see benefit of Al to detect cancer risk

75% concerned about driverless cars

People can see both the benefits and risks

Even though majority see benefits of Al in cancer detection, 64% still worried about relying too heavily on technology rather than professional judgements

Concern is rising

59% are concerned about the use of Al to determine eligibility for welfare benefits such as Universal Credit. This figure was 44% in 2022/23

Awareness and use of LLMs is rising rapidly

Gaia Marcus@gaiamarcus· Mar 26

Now hearing from @adalovelaceinst.bsky.social's Ellie on her participatory research on the intersections between AI and their ambitions for the public good www.adalovelaceinstitute.org/report/ai-pu...

Slide that reads "what do diverse ppl in different areas think about the public good

Publics hold composite views, expecting social changes but unsure what it might mean for them

Ai for public good means, careful AI use that benefits everyone and confronts society's big challenges

Gaia Marcus@gaiamarcus· Mar 26

Now UNJUST - who are challenging the data (eg the old gangs violence matrix) used to profile young black boys and men in London

Gaia Marcus@gaiamarcus· Mar 26

Now we're hearing from an art's based organisation and their research

Gaia Marcus@gaiamarcus· Mar 26

At a very packed - standing room only - event in parliament Public Voices in AI event. Heading from the worker's observatory first

Gaia Marcus@gaiamarcus· Feb 6

First @adalovelaceinst.bsky.social presentation as part of AI action summit warm-up here in paris - @agstrait.bsky.social setting out background, contrctyand importance of audits in ensuring that AI systems work, work as expected and are understood.

Gaia Marcus@gaiamarcus· Jan 8

Happy new year!! 🎊🎊 The @adalovelaceinst.bsky.social team continue to have excellent taste in gifts... Joy, hope and some @anthonyburrill.bsky.social flair

Gaia Marcus@gaiamarcus· Dec 10

Well that's the new year's resolution sorted! A very joyful secret Santa present via @adalovelaceinst.bsky.social colleagues

Gaia Marcus@gaiamarcus· Dec 5

@imogen-parker.bsky.social this chimes with some of the arguments you had on consultations in the @jrf-uk.bsky.social piece

Extract from JRF article by Imogen Parker "Replacing existing consultation analysis with this Consult tool would save thousands of hours of work. However, this approach privileges quantitative over qualitative analysis, undermining the potential of finding a ‘golden nugget’ of data: a compelling quote or story that shifts thinking. It would also erode the experience of engaging with consultation material, dulling the cumulative impact of reading people’s stories in their voices. Sometimes ‘friction’ in a system can produce something positive, intimate, or tactile which gets erased through the pursuit of technological efficiency.

Of course, it’s idealised to imagine those designing policy have free reign to be moved by individual responses in consultations. The reality is often a regimented coding process. But automating engaging with the material directly takes away one important aspect of a consultation, to offer publics a chance to be listened to.

No one would argue that consultation fosters a perfect human relationship with a government decision-maker. It is, in part, an imperfect attempt at something relational. We should consider how automating such processes may change the willingness of the public to engage, and how they choose to do so, if they know their responses are to be analysed by a machine. We need to count the value given to existing human-to-human relationships."

Gaia Marcus@gaiamarcus· Dec 5

On my panel today I covered: 1. Much of the public debate on AI is based on the assumption that its impact on our society and economy will be transformational. If we accept this at face value, we should expect its governance to reflect that of similarly consequential systems: this means action.

Panel at 'No AI without women' featuring me, Asma Mhalla and Ysens de France
Photo via Alexandra Alberti-Verdenne

Flyer about No Ai Without Women, featuring all the speakers

Gaia Marcus@gaiamarcus· Dec 5

Really looking forward to hearing from @sashamtl.bsky.social on sustainability and AI

Panel at 'No AI without women' event, featuring Dr Sasha Luccioni - AI & Climate lead at Hugging Face. Really baroque room and furnishings at the Elysée Palace!

Gaia Marcus@gaiamarcus· Dec 5

Delighted to be in the Elysée palace for the 'No AI without women' event that's part of the run-up to February's AI action summit. Fantastic panelists and topics

Gaia Marcus@gaiamarcus· Nov 28

It's the best time of the ... *checks notes* fortnight! @adalovelaceinst.bsky.social newsletter time. But no fear - sign up for this JOY in your inbox, too. Report updates, jobs, events, reading lists, oh my.. nuffieldfoundation.tfaforms.net/149

Gaia Marcus@gaiamarcus· Nov 18

'Aside from the Nobel Laureate...' Hannah Fry @fryrsquared.bsky.social and Demis Hassabis in conversation

Closing Fireside Sir Demis Hassabis, Co-Founder & CEO, Google DeepMind In Conversation with Hannah Fry

Gaia Marcus@gaiamarcus· Nov 18

Really appreciating hearing from @aaberhe.com Asmeret Asefaw Berhe, who is calling for transformative thinking around the next generation of computing in terms of sustainability. Underscores that incremental thinking will not be enough in terms of the real-world energy costs/usage on technology

Paul Hofheinz, President and Co-Founder, Lisbon Council, in conversation with: Asmeret Asefaw Berhe, Professor of Soil Biogeochemistry and Falasco Chair in Earth Sciences, University of California, Merced Fabian J. Theis, Director of the Institute of Computational Biology and Professor at TUM Mathematics & Life Sciences Bosun Tijani, Minister of Communications, Innovation and Digital Economy, The Federal Republic of Nigeria

Gaia Marcus@gaiamarcus· Nov 18

Really impressed by Alison Noble - bringing up a lot of the sociotechnical questions surrounding use of AI in science and medical fields. Understanding that change in technology changes relational aspect and how humans in wider system act and operate.