Rebecca Anselmetti

🇬🇧 UK AI Policy Update: Evolving Approach to AI Governance ⭕ Recent developments suggest the UK is refining its approach to AI regulation: 🔴 UK signs Council of Europe's AI Treaty: An international legally binding agreement focusing on democracy and human rights in AI. Signed last week, it signals the UK's commitment to ethical AI and international cooperation. 🔴 Lord Tim Clement-Jones' AI Bill: The Public Authority Algorithmic and Automated Decision-Making Systems Bill had its first reading in Parliament yesterday. While it’s a private member’s bill, it reflects growing interest in stronger AI governance. Full text https://lnkd.in/eqd3qVuz These steps indicate a shift from last year's focus on existential risks and the light-touch, pro-innovation approach. #AIPolicy #UKTech #AIEthics #AIGovernance

13

1 Comment

From a fantastic piece by Iain Nash and Pier Giorgio Chiara, Ph.D.: 'Despite receiving scant media and academic coverage and discussion, ... there are few technology products for sale within the European Union that the CRA will not impact. 'The CRA [art 57] mandates that products with digital elements ... must not pose risks to the health or safety of individuals, or conflict with obligations under union or member states laws designed to protect fundamental rights. '... these assessments ... effectively making them normative judgments ... evidence of fundamental rights violations used in these assessments is not confined to the European Union. '... manufacturers of specialized processors, sensors, or algorithms could find that all products utilizing these components might be removed from the European Union due to their use in other jurisdictions. 'If the European Commission intended to develop an analogue to the Alien Torts Statute, which allows the removal of products from the European Market (as opposed to enabling a claim to be made against the technology company itself), this should be in a separate and distinct piece of legislation and not delivered as part of a cybersecurity regulation'. https://lnkd.in/gGsttsuB

1

Make sure you provide your opinion on risk thresholds for advanced AI systems. We need everybody to weigh in. Deadline is September 10th. To me, compute threshold are not a useful metric for identifying safety risks. AI safety is a context-dependent evaluation of multiple factors, not a distinct property of a model. Also, passing a compute threshold does not necessarily indicate the presence of dangerous capabilities. In fact, it may well be that more compute will help in achieving higher levels of safety. Also, recently greater levels of performance are being achieved with smaller and smaller models. Rather than setting thresholds (based on compute or other), the evaluation of capabilities and, even more importantly, limitations of AI systems is a much better indication of issues that can turn in possible risks when the model will be used. Work on evaluation is ongoing in both academia and corporate environments, and there is still not a single unique way to do it. However, the danger in running a certain risk is very related to the deployment and use scenario. Even model with powerful capabilities or serious limitations can be safe to use in certain scenarios but dangerous in others. What is your opinion? Whatever it is, you should upload it to the OECD.AI site for the public consultation (https://lnkd.in/dKDP_4za). I already uploaded mine! #ai #airisks #advancedai #computethresholds

120

12 Comments

THE CYBER RESILIENCE ACT HAS BEEN ADOPTED BY THE COUNCIL OF THE EUROPEAN UNION! The EU is weeks away from becoming the first jurisdiction with a bespoke regulatory framework for the product security _and_ labelling of all software sold commercially in the EU (save stuff covered by other EU rules like cars and healthtech). Yes, the Yanks (via EO14028–>NIST) defined critical software (the CRA has ‘important products with digital elements’ and ‘critical products with digital elements’), but the Yanks, for now at least, have only gone down the procurement route for regulating vendor SDLCs. The EU, on the other hand, is covering everything sold commercially (bar the stated exceptions) to anyone in the EU. Big day for all us SDLC regulation people! What happens next: Council and EuroParl President sign it —> Publication in the EU OJ —> Entry into force 20 days later —> Application of most provisions 36 months later. Press release (includes link to final text): https://lnkd.in/gJT3Aqxi

3

UK Parliamentary Report on AI Governance. It includes most of the things from the deployment of AI to principles and legislations for its usage, the effects of it on sectirs and in general, institutional control mechanisms and so on. It is better to read it.

1

Good regulation, well resourced, communicative, educating...and with teeth when needed.

1

1 Comment

As #artificialintelligence usage continues its explosive growth, the United States Congress and state legislators alike have sprung into action, proposing and adopting various new pieces of legislation aimed at safeguarding individuals’ rights against so-called “digital replicas.” In a September 18 webinar, Sarah Moses, Nathaniel Bach and Paul Luehr will discuss these developments, prospects for passage of pending bills, and how companies and artists can adapt to new changes in the law and ensure their rights are protected. Learn more and register here: https://lnkd.in/e_WP8Mm7

9

PUBLIC CONSULTATION ON RISK ⚠️ THRESHOLDS https://lnkd.in/e98Pzw-b The OECD is joining forces with a diverse array of stakeholders to explore potential approaches, opportunities and limitations for establishing risk thresholds for advanced AI systems. To inform this work, we are holding an open public consultation to obtain the views of all interested parties. We are interested in hearing your thoughts on the following key questions: ❓ What publications or other resources have you found useful on the topic of AI risk thresholds? ❓ To what extent do you believe AI risk thresholds based on compute power are adequate and appropriate to mitigate risks from advanced AI systems? ❓ To what extent do you believe that other types of AI risk thresholds would be valuable, and what are they? ❓ What strategies and approaches can governments or companies use to identify and set out specific thresholds and measure real-world systems against those thresholds? What requirements should be imposed for systems that exceed any given threshold? ❓ What else should the OECD and collaborating organisations keep in mind with regards to designing and/or implementing AI risk thresholds? 10 SEPTEMBER DEADLINE TO PARTICIPATE https://lnkd.in/e98Pzw-b Francesca Rossi Stuart Russell Michael Schönstein Ulrik Vestergaard Knudsen Jerry Sheehan Audrey Plonk Celine Caira Luis Aranda Jamie Berryhill Lucia Russo Noah Oder John Leo Tarver ⒿⓁⓉ Rashad Abelson Angélina Gentaz Valéria Silva Bénédicte Rispal Johannes Leon Kirnberger Eunseo Dana Choi Pablo Gomez Ayerbe Sara Fialho Esposito Nikolas S. Sarah Bérubé Guillermo H. #airisk #aisafety #trustworthyai #oecd #risk

84

2 Comments

Thanks Giulio Zucchini for sharing on Domani some great reflections from the recent - excellent - EU DisinfoLab conference and for mentioning the work we CyberPeace Institute are doing when it comes to #disinformation #cybersecurity #AI and its impact on the #threat landscape, when it comes to #maliciousactors evolution, the #harm on #civilsociety and the impact on #civicspace #democraticspace More here: https://lnkd.in/d3ubx8ti #disinformazione #socialmedia #fakenews #cybersicurezza #AI #societacivile

20

1 Comment

I imagine this will have to be said over and over and over again until policy-makers get it: AI will not solve the challenges facing the public sector! The only solution is a radical shift towards a state and public sector that does things *with* individuals, families and communities rather than doing things *to* them. Tech has a part to play in that but as a tool for empowerment and agency not as a solution in itself. If we simply see AI as a way to process people faster through a bureaucratic system, it will just embed the do-to mindset and all the problems it causes even more deeply. #NHS #AI #tech #localgov #publicservices #productivity #innovation https://lnkd.in/e5MZR3gd

306

53 Comments

Ahead of the upcoming ISO/TC307 Plenary regarding standards for #Blockchain & #DistributedLedgerTechnology [1], I wonder: - Are there enough completed standards for DLT Interoperability? INATBA's comprehensive review of standards [2] suggests not. In terms of upcoming technical blockchain interoperability standards, ones to watch are: - ISO TC307's Interoperability Framework [3] - IETF's Secure Asset Transfer Protocol (SATP) [4]. - EEA's Distributed Ledger Interoperability Technical Specification [5] ISO TC307's Interoperability Framework standard will conceptualise and categorise the interoperability domain in a clear manner, including by providing a reference architecture. A key source text of this standard is [6] (ACM's 2023 most-read paper on the topic of DLT [7]). Regarding the advantages of SATP, they are discussed in Sections 1, 2, 3 & 5 of [8]. I am also looking to review the EEA's draft specification, and hoping that it can be more widely applicable than just to EVM based blockchains. References: [1] https://lnkd.in/eAW_uHaU [2] https://lnkd.in/ezWBpXFZ [3] https://lnkd.in/eza3FfkZ [4] https://lnkd.in/eAxrTk9u [5] https://lnkd.in/eGXcKeWV [6] https://lnkd.in/evbUh2s7 [7] https://lnkd.in/e6sNf-qr [8] https://lnkd.in/eazNg6Xc

46

3 Comments

DORA Policy Updates Released 🔐 🇪🇺 The European Supervisory Authorities (ESAs) have released the second batch of policy products under the Digital Operational Resilience Act (DORA). This includes four regulatory technical standards, one implementing technical standard, and two guidelines aimed at improving the digital operational resilience of the EU's financial sector. Key areas covered include reporting frameworks for ICT-related incidents and threat-led penetration testing, as well as guidelines for oversight cooperation. The new standards and guidelines are now with the European Commission for review and adoption. View the link for more information 🔗 https://lnkd.in/eeaKNSXr #DORA #DigitalResilience #FinancialSecurity #RegulatoryUpdates #Fintech #ESAs #ICTSecurity #Cybersecurity #OperationalResilience #EURegulations

2

LAST FEW DAYS TO PARTICIPATE! 📅 DEADLINE: 1 OCTOBER Public consultation on risk thresholds for advanced AI systems 👉 https://lnkd.in/e98Pzw-b The OECD is working with diverse stakeholders to explore potential approaches, opportunities, and limitations for establishing risk thresholds for advanced AI systems. To inform this work, we are holding an open public consultation to get the views of all interested parties. We are interested in hearing your thoughts on the following key questions: ❓ What publications or other resources have you found helpful on AI risk thresholds? ❓ To what extent do you believe AI risk thresholds based on compute power are adequate and appropriate to mitigate risks from advanced AI systems? ❓ To what extent do you believe other AI risk thresholds would be valuable, and what are they? ❓ What strategies and approaches can governments or companies use to identify and set specific thresholds and measure real-world systems against those thresholds? What requirements should be imposed for systems that exceed any given threshold? ❓ What else should the OECD and collaborating organisations consider concerning designing and/or implementing AI risk thresholds? 📅 10 SEPTEMBER DEADLINE TO PARTICIPATE 👉 https://lnkd.in/e98Pzw-b Francesca Rossi Stuart Russell Michael Schönstein Ulrik Vestergaard Knudsen Jerry Sheehan Audrey Plonk Celine Caira Luis Aranda Jamie Berryhill Lucia Russo Noah Oder John Leo Tarver ⒿⓁⓉ Rashad Abelson Angélina Gentaz Valéria Silva Bénédicte Rispal Johannes Leon Kirnberger Eunseo Dana Choi Pablo Gomez Ayerbe Sara Fialho Esposito Nikolas S. Sarah Bérubé Guillermo H. #airisk #aisafety #trustworthyai #oecd #risk

25

2 Comments

Great to see techUK's Quantum Resilience Industry Perspective covered in POLITICO yesterday (extra points for the very Taylor Swift coded sub heading too) In case you missed it: Following the long-awaited guidance from NIST in August 2024, techUK's new paper sets out practical advice on cryptographic solutions such as post-quantum cryptography (PQC), and quantum-based solutions such as quantum key distribution (QKD). Through discussing these in simple terms, we hope to show that the pathway to quantum resilience doesn't have to be daunting. Quantum technologies bring benefits but also pose security risks. This report provides insights to guide businesses through these challenges. You can read more here ➡ https://lnkd.in/exNZvg3G Huge thanks to all of the quantum and cryptography experts across techUK's membership who shaped this report. On the techUK side, thanks to Jill Broom, Claire Leslie and Margherita Certo for getting this over the line 💪 #quantum #security #quantumresilience #PQC #QKD

23

2 Comments

Let's talk about AI in Policing AI presents ethical dilemmas for policing but not barriers to adoption. We  recognise the emergence and value of ethical and scrutiny arrangements for  AI in policing including for example, The Independent Digital Ethics Panel for  Policing and the government’s Centre for Data Ethics and Innovation. Indeed, the extent to which artificial intelligences ( AIs ) can influence policing  is a topic which is seeing greater momentum and debate, including the  surrounding ethical issues. However, perhaps it is fair to say that AI is a term  which is now commonly used but less understood by the policing community.  Arguably, a strong partnership is needed between policing and the technology  industry to understand how AI can improve the efficiency and effectiveness of  policing. Interested in more information? Please check out this link for access to our full white paper on AI in Policing ~ https://lnkd.in/ew6PsJyC #aiinpolicing #aipolice #innovation #responsibleai

2

New commentary from Natasha Buckley and Neil Ashdown on the relationship between technology innovation, national security and geopolitics. In an era of great power competition when it comes to developing technology, national security and resilience matters. Read more 👇

10

There are already over 150 research managers from across the globe signed up to our "Promoting European cooperation in developing responses to research security challenges: Online Events Series"!! 🌍 In case you missed it, link to book your place in the comments 👇🏽 Operationalising “research security” measures has become a defining challenge for research organisations and is one that disproportionately affects research managers as the need for enhanced due diligence and engagement with regulatory compliance increases. In January 2024, the European Commission Council Recommendation on Research Security highlighted the need to enhance research security across Europe and outlined the creation of a European centre of expertise on research security, the development of tools to support research organisations and consistency of approach, as well as plans to change the EU’s dual-use export regime. Building on the foundations of the #ResearchSecuritySummit at the EARMA Annual Conference 2024 in Odense, Denmark, the aim of this online event series is to address the needs of the research management community in responding to research security. The three free to attend online events focus on the following research security areas most relevant to European research managers: 1) Research security risks: insider threats (October 21st 2024) 2) Resources for managing research security risks (November 15th 2024) 3) Products to support research security management (December 12th 2024) Each session will provide practical support to research managers including discussion of case studies, signposting to publicly available tools and resources, and commercially available products. #Research #Innovation #ResearchSecurity #TrustedResearch #SecureInnovation #ResearchManagement Jennifer Johnson Joe Timlin Linsey Dickson Duncan Jackson Chaitali Desai

15

3 Comments

The #OneLogin team in the Government Digital Service is working on a Shared Signals Framework as part of the effort to identify #fraudulent behaviour in using government services - Dr Louise Maynard-Atem, deputy director, data and insights at GDS, described the plan in her presentation to the recent UKAuthority #AI and #Data4Good conference. Maynard-Atem said. “When we are seeing indications of potentially fraudulent behaviour, we can share that with other services aligned to One Login so they can be aware of any risks at front door and potentially play it through to the interactions the system has with the service." #interoperability #APIs #fraud #ProtectingThePublicPurse https://lnkd.in/e27X7Hu8

8