Lightnews — Scholar-powered news

The Alan Turing Institute @theturing.bsky.social · 9h

🚨 Published today: First International AI Safety Report update

➡️ Focus on capability advancements since the last report & implications for risks.

➡️ 100+ experts contributed, supported by 30 countries, the EU, OECD & UN.

➡️Read the update: internationalaisafetyreport.org/publication/...

1

The Alan Turing Institute @theturing.bsky.social · 1d

🌐 California’s new law on frontier models sets a precedent for developers at the heart of the global AI industry.

In a new CETaS Expert Analysis, Connor Attridge looks at how the UK can build on the law as it navigates the relationship between AI and national security:
https://bit.ly/493YXsV

California’s Frontier AI Act: National Security Implications for the UK

bit.ly

The Alan Turing Institute @theturing.bsky.social · 1d

❓ Are you a researcher looking to make real-world impact as part of an expert multi-disciplinary team?

Apply to join the Turing’s January #DataStudyGroup: Modelling AI competency needs in the UK labour market

📅26 Jan – 6 Feb 2026
➡️ Apply now:https://bit.ly/42EKbF3

Slide shows people working together in a DSG with quotes: "I’m actually really excited to hopefully come back to one of these", “I can’t believe the amount that I’ve learned this week”

Slide shows two quotes: "It’s really nice to have different levels of experience to work with", "It was great to see how much could be achieved in just a span of five days"

Slide shows a woman presenting her work with this quote: “It was great to see… how efficient it is to work in a multidisciplinary team because that’s not something you get to do very often when you’re in a PhD”

The Alan Turing Institute @theturing.bsky.social · 2d

🔐 Securing Critical National Infrastructure in a Data-Driven World

Read our new report, in partnership with @lloydsregister.bsky.social, introducing a roadmap to use data for more secure, resilient & sustainable infrastructure: https://bit.ly/4q98B3B

Deep Dive on Data and Information for Critical Infrastructure Management and Maintenance

This report examines the role of data and information in managing and maintaining critical infrastructure in the UK.

bit.ly

The Alan Turing Institute @theturing.bsky.social · 6d

As model capabilities increase, more work to defend against data poisoning is essential to ensure their secure and trustworthy deployment.

➡️Read the study: arxiv.org/abs/2510.07192

Poisoning Attacks on LLMs Require a Near-constant Number of Poison Samples

Poisoning attacks can compromise the safety of large language models (LLMs) by injecting malicious documents into their training data. Existing work has studied pretraining poisoning assuming adversar...

arxiv.org

The Alan Turing Institute @theturing.bsky.social · 6d

⚠️ Data poisoning occurs when individuals distribute online content designed to corrupt an AI model’s training data, producing potentially harmful outputs.

🚪 This includes the insertion of backdoors – specific phrases that trigger an otherwise-hidden behaviour.

1

The Alan Turing Institute @theturing.bsky.social · 6d

Researchers from the Turing, Anthropic and the AI Security Institute have conducted the largest study of data poisoning to date.

Results show that as little as 250 malicious documents can be used to “poison” a language model, even as model size and training data grow: bit.ly/4n0mH4t

LLMs may be more vulnerable to data poisoning than we thought

A Turing collaboration with the AI Security Institute and Anthropic will help to protect AI models from misuse

bit.ly

1 3 4

The Alan Turing Institute @theturing.bsky.social · 6d

Defence AI Assurance: Identifying promising practice and a system card template for Defence

❓ How can the MoD’s AI policy be implemented in real-world operations?

Read the latest report from Turing Senior Research Associate, Anna Knack and co-authors: https://bit.ly/48mUMIt

Defence AI Assurance: Identifying Promising Practice and A System Card Template for Defence

The Alan Turing Institute conducted this study in collaboration with Accenture and in coordination with the Ministry of Defence’s Defence AI and Autonomy Uni

bit.ly

The Alan Turing Institute @theturing.bsky.social · 7d

💻 Children are facing unprecedented exposure to harmful content online

💡 In a new CETaS Briefing Paper, Sam Stockwell & Rosamund Powell explore how the 🇬🇧, 🇺🇸, 🇪🇺 & 🇦🇺 are using #AgeAssurance tech to tackle this problem, unpacking limits & lessons.

http://bit.ly/4765PDQ

Age Assurance Technologies and Online Safety

bit.ly

The Alan Turing Institute @theturing.bsky.social · 7d

To ensure our qualitative research is as open as possible, we’ve taken a reproducible approach by preregistering our study design.

➡️ Access the preregistration here: osf.io/d9mgb

OSF

osf.io

The Alan Turing Institute @theturing.bsky.social · 7d

🔎 New qualitative preregistration published: Advancing Biomedical Data Science Careers

In collaboration with @ebi.embl.org, we’re addressing the urgent need in biomedical research to better understand data science careers.

Learn more about the project: bit.ly/4o0Vdg6

Advancing Biomedical Data Science Careers

Conferences, workshops, and other events from around the Turing Network

bit.ly

1

The Alan Turing Institute @theturing.bsky.social · 8d

📢 Survey call: Data-Centric Engineering community

Do you work at the intersection of AI, data science and engineering?

💡 Share your career path, diversity and role insights in our survey: https://bit.ly/48RGRua

#DataCentricEngineering #DCE

Data-Centric Engineering community survey

Share your experience. We are interested in learning more about the people that make up our community.

bit.ly

The Alan Turing Institute @theturing.bsky.social · 9d

How is AI changing our lives?

🎧 The Turing's Adrian Weller joined Brian Cox, Neil Lawrence, Steph Wright and Jeanette Winterson to explore this fundamental question.

Listen to the latest episode of @crick.ac.uk ’s A Question of Science podcast: https://www.lnk.to/AQOSAISH

Group photo on podcast stage with Neil Lawrence, Steph Wright, Brian Cox, Jeanette Winterson and Adrian Weller

The Alan Turing Institute @theturing.bsky.social · 12d

Explore the country profiles published so far:

🇨🇦 Canada: bit.ly/4o5iHko

🇪🇺 European Union: bit.ly/4mFw89n

🇮🇳 India: bit.ly/42R2jM4

🇸🇬 Singapore: bit.ly/4gPiqPU

The Alan Turing Institute @theturing.bsky.social · 12d

This project will explore:
✅ Shifting national priorities on AI
✅ Dynamics of cooperation & competition
✅ Potential areas of international alignment

1 2

The Alan Turing Institute @theturing.bsky.social · 12d

🌍What does the future hold for international AI governance?

As the world balances competition with international cooperation & safety, we’re mapping the evolving landscape of AI regulation & standardisation through a series of country profiles: https://bit.ly/4pQiQK4

AI governance around the world

bit.ly

1

The Alan Turing Institute @theturing.bsky.social · 13d

🌐 Enhancing global AI governance

As part of our work w/ the FCDO we’re in Hanoi this week hosting a multi-day workshop on AI governance & ethics.

Supporting a responsible environment for tech development is key to protecting those at risk while enabling innovation to thrive.

The Alan Turing Institute @theturing.bsky.social · 14d

👥 Persona agents are widely used in roles such as AI assistants & NPCs, but to what extent do they replicate the richness of human behaviour?

A new CETaS Expert Analysis explores how psychometrics can – and can’t – provide insights into these agents: https://bit.ly/4o0cEh0

Patterns, Not People: Personality Structures in LLM-powered Persona Agents

bit.ly

1

The Alan Turing Institute @theturing.bsky.social · 14d

💡 Want to apply your data skills to real-world challenges?

Apply to work on challenges with real-world datasets from the Alan Turing Institute in collaboration with innovateuk.bsky.social #BridgeAI (supported by PwC).

📅26 Jan – 6 Feb
⏰Applications close: 26 Oct
➡️https://bit.ly/42EKbF3

The Alan Turing Institute @theturing.bsky.social · 16d

🔎 Examining AI assurance from a UK Defence perspective

New research published today with the Ministry of Defence, Defence AI Centre and Accenture, explores the assurance of safe and responsible AI in defence, considering real-world military use cases.

➡️Learn more: http://bit.ly/470EF24

New research paves way for safe and responsible AI in defence

The Alan Turing Institute in coordination with the Ministry of Defence (MoD) and

bit.ly

The Alan Turing Institute @theturing.bsky.social · 18d

AI for Decarbonisation: Building Sustainable AI webinar

📅6 Oct, 11:30–12:30
➡️Register: bit.ly/4gHcWXv

Hear from experts covering practical pathways to decarbonising the AI supply chain:

✅E-waste & dark data
✅Efficient algorithms & data architectures
✅Next-gen hardware

ADViCE Webinar Series: Building Sustainable AI

Who should attend? This event will be of interest to professionals and stakeholders across public sector, AI, net zero, cleantech, energy, tran

bit.ly

The Alan Turing Institute @theturing.bsky.social · 20d

#TuringConnections PhD Student Presentations & Networking Day

📅24 Oct 2025, 10:00–17:00
📍 @ucl.ac.uk
➡️Register by 15 Oct 2025: bit.ly/4pAaNAV

🌟 #DataScience & #AI PhD students: present research, develop skills & connect with peers at this free event.

#PhDNetworking

PhD Student Presentations and Networking Day - October 2025

Conferences, workshops, and other events from around the Turing Network

bit.ly

1

The Alan Turing Institute @theturing.bsky.social · 21d

📢 Applications open: #DataStudyGroup

The challenge: Modelling AI competency needs in the UK labour market

📅 26 Jan – 6 Feb 2026
➡️ Apply now: bit.ly/42EKbF3

Challenge set by the Alan Turing Institute in collaboration with @innovateuk.bsky.social #BridgeAI (supported by PwC).

Data Study Group - January 2026

What are Data Study Groups?

bit.ly

1

The Alan Turing Institute @theturing.bsky.social · 23d

What do UK citizens really think about AI-powered data processing for national security?

Join CETaS' UK Public Attitudes to National Security Data Processing webinar to explore insights from their novel research.

📅 2 Oct, 15:00–16:00 📍Online
➡️ bit.ly/3VtmJXr

CETaS Webinar: UK Public Attitudes to National Security Data Processing

As AI and machine learning transform the way national security agencies process vast amounts of information, they bring both crucial opportunities and seriou

bit.ly

1

The Alan Turing Institute @theturing.bsky.social · 28d

We’re pleased to be at #LISW2025 hosting a breakfast briefing this morning to explore AI and defence for maritime.

Bringing together leaders across the global maritime sector, we'll examine AI applications and vulnerabilities, and how attackers can use AI against maritime technology.

Kimberly Tam from the Alan Turing Institute presenting at London International Shipping Week 2025

2