About
After completing a passion project in 2009, writing and directing Wales’s first online…
Articles by James
-
Data Protection Perspective #1 - “you don’t appreciate the bombs that don’t go off”.
Data Protection Perspective #1 - “you don’t appreciate the bombs that don’t go off”.
I’ve been accused in the past of being able to make a “boring” subject fascinating so maybe it’s no wonder I ended up…
-
GDPR – COMMON MISCONCEPTIONS
GDPR – COMMON MISCONCEPTIONS
Where do you get your GDPR information from? Legal advisers? Data protection practitioners? LinkedIn? Or maybe from…
3 Comments
Activity
12K followers
-
James Robson posted thisChildren are not failing to manage these platforms. The platforms are getting very good at managing children. That was the real story in Parliament this week when Meta, TikTok and Roblox were pressed on addictive design and child safety. I do not see this as a narrow moderation debate. I see it as a data governance issue. These services learn from behaviour at industrial scale. A pause. A replay. A late-night scroll. A burst of likes on a certain kind of content. Those signals are logged, ranked and fed back into systems designed to hold attention. That matters more when the user is 13 than when the user is 33. In the UK, we often split this conversation in two. Safety sits in one box. Privacy sits in another. That split no longer works. If a platform is constantly profiling a child in order to shape what they see next, the governance of that data is part of the safety question. It is not background detail. It is the mechanism. That is why this moment matters. Parliament is not just challenging a few product choices. It is pushing on a deeper issue. How much behavioural data should platforms be allowed to use when the outcome is stickier, more compulsive use by children? I am not anti-technology. I work in data protection because I believe good data use can improve lives. But this is where the line needs to be drawn with care, and with some backbone. The UK has a chance to be practical here. Less theatre. More clarity on profiling, age assurance, default settings and accountability for recommender systems. I’d be interested in where others would draw the line between personalisation and manipulation when children are involved. #DataGovernance #ChildSafety #Privacy #Data #DataSharing
-
James Robson reposted this𝐈’𝐦 𝐬𝐩𝐞𝐚𝐤𝐢𝐧𝐠 𝐚𝐭 𝐒𝐩𝐚𝐧 𝐂𝐲𝐛𝐞𝐫 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐀𝐫𝐞𝐧𝐚 𝐢𝐧 𝐏𝐨𝐫𝐞č 𝐨𝐧 22 𝐌𝐚𝐲 2026. My session, Data Protection in the UK General Election, will look at the privacy, governance, and cyber security issues that come with handling election-related data. Looking forward to the discussion and to meeting others at the event. #DataProtection #Privacy #InfoSec #SpanCyberSecurityArena
-
James Robson reposted thisI think we are underestimating the most dangerous AI risk in finance. It is not what one model gets wrong. It is what happens when everyone starts using the same model, in the same way, at the same time. That is why the Bank of England’s latest position matters. Its April 2026 assessment was measured, but important. The Bank said advanced AI is not yet being used in finance in a way that creates systemic risk, but it also warned that the risk could rise quickly as adoption deepens, especially around agentic AI, operational dependence, and cyber exposure. Andrew Bailey’s warning about cyber risks linked to Anthropic’s latest model pushed that point further. For me, that is the real signal. AI governance is moving out of ethics decks and into resilience, market stability, and concentration risk. The valuable question for UK firms is no longer just, “Is this tool accurate?” It is: ~ Who else relies on the same provider? ~ What fails if that provider fails? ~ How quickly would we detect shared model-driven behaviour across critical decisions? That shift matters because most serious failures in finance are not isolated. They spread through correlation, dependency, and speed. So this feels like a leadership moment. Privacy, security, risk, and governance teams should not be cast as the people who slow AI down. They should be the people who make it safe enough to scale. That means stress-testing concentration risk, mapping third-party AI dependencies, and treating model adoption as an operational resilience issue from day one. I am curious whether others are seeing the same change. Are your AI conversations still focused on capability, or are they finally getting serious about dependency and systemic risk? #AI #UK #Finance #Data #DataSharing
-
James Robson posted thisI think we are underestimating the most dangerous AI risk in finance. It is not what one model gets wrong. It is what happens when everyone starts using the same model, in the same way, at the same time. That is why the Bank of England’s latest position matters. Its April 2026 assessment was measured, but important. The Bank said advanced AI is not yet being used in finance in a way that creates systemic risk, but it also warned that the risk could rise quickly as adoption deepens, especially around agentic AI, operational dependence, and cyber exposure. Andrew Bailey’s warning about cyber risks linked to Anthropic’s latest model pushed that point further. For me, that is the real signal. AI governance is moving out of ethics decks and into resilience, market stability, and concentration risk. The valuable question for UK firms is no longer just, “Is this tool accurate?” It is: ~ Who else relies on the same provider? ~ What fails if that provider fails? ~ How quickly would we detect shared model-driven behaviour across critical decisions? That shift matters because most serious failures in finance are not isolated. They spread through correlation, dependency, and speed. So this feels like a leadership moment. Privacy, security, risk, and governance teams should not be cast as the people who slow AI down. They should be the people who make it safe enough to scale. That means stress-testing concentration risk, mapping third-party AI dependencies, and treating model adoption as an operational resilience issue from day one. I am curious whether others are seeing the same change. Are your AI conversations still focused on capability, or are they finally getting serious about dependency and systemic risk? #AI #UK #Finance #Data #DataSharing
-
James Robson shared this𝐈’𝐦 𝐬𝐩𝐞𝐚𝐤𝐢𝐧𝐠 𝐚𝐭 𝐒𝐩𝐚𝐧 𝐂𝐲𝐛𝐞𝐫 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐀𝐫𝐞𝐧𝐚 𝐢𝐧 𝐏𝐨𝐫𝐞č 𝐨𝐧 22 𝐌𝐚𝐲 2026. My session, Data Protection in the UK General Election, will look at the privacy, governance, and cyber security issues that come with handling election-related data. Looking forward to the discussion and to meeting others at the event. #DataProtection #Privacy #InfoSec #SpanCyberSecurityArena
-
James Robson reposted thisWhat if the next big step forward for privacy does NOT come from more notices, more pop-ups, or more settings? What if it comes from better AI? That sounds counterintuitive at first. AI is usually framed as the privacy challenge. And often, rightly so. It can infer more, move faster, and act with less friction than most organisations are ready for. But there is another side to the conversation. The real privacy problem in many organisations is no longer a lack of principles. It is a lack of capacity. Too many systems. Too many vendors. Too many data flows. Too many moments where people are expected to understand, question, and manage complexity that no normal person can realistically keep up with. So perhaps the more useful question is this: Can AI help restore human control over data, rather than erode it? Used well, it could help people understand how decisions are made, surface hidden risks earlier, make privacy choices easier to find, and reduce the burden on individuals who are still expected to navigate dense policies and fragmented controls on their own. That does not reduce the need for governance. It makes good governance more important. And it feels especially timely in the UK, where the real opportunity may be to build AI that supports trust, accountability, and better data sharing. Not just faster processing. Maybe the future of privacy is not less technology. Maybe it is technology finally doing some of the hard work for people. What would need to be true in your organisation for AI to make privacy simpler, fairer, and more human? #DataSharing #Data #DataProtection #Privacy #UK
-
James Robson shared thisWhat if the next big step forward for privacy does NOT come from more notices, more pop-ups, or more settings? What if it comes from better AI? That sounds counterintuitive at first. AI is usually framed as the privacy challenge. And often, rightly so. It can infer more, move faster, and act with less friction than most organisations are ready for. But there is another side to the conversation. The real privacy problem in many organisations is no longer a lack of principles. It is a lack of capacity. Too many systems. Too many vendors. Too many data flows. Too many moments where people are expected to understand, question, and manage complexity that no normal person can realistically keep up with. So perhaps the more useful question is this: Can AI help restore human control over data, rather than erode it? Used well, it could help people understand how decisions are made, surface hidden risks earlier, make privacy choices easier to find, and reduce the burden on individuals who are still expected to navigate dense policies and fragmented controls on their own. That does not reduce the need for governance. It makes good governance more important. And it feels especially timely in the UK, where the real opportunity may be to build AI that supports trust, accountability, and better data sharing. Not just faster processing. Maybe the future of privacy is not less technology. Maybe it is technology finally doing some of the hard work for people. What would need to be true in your organisation for AI to make privacy simpler, fairer, and more human? #DataSharing #Data #DataProtection #Privacy #UK
-
James Robson reposted thisPrivacy notices will not fix vendor lock-in. A lot of UK privacy work still focuses on the visible layer: notices, retention periods, DPIAs, data-sharing clauses. Necessary work. But it misses something bigger. If the systems underneath public services, healthcare, education, and government are heavily dependent on a small number of overseas providers, privacy stops being just a compliance exercise. It becomes a question of control, resilience, and public trust. That is why the digital sovereignty debate matters. The point is not to retreat from innovation, or pretend every global supplier is a problem. The point is to be honest about dependency. When core infrastructure is hard to inspect, hard to exit, and hard to challenge, risk builds quietly. Procurement teams feel it. Security teams feel it. Privacy teams should be talking about it far more than we do. The attached report makes a strong case that the UK has looked past this for too long. It links digital dependence to security exposure, weak bargaining power, poorer public value, and less room for democratic choice. It also argues that open standards, open source, and shared digital infrastructure offer a more durable path. That feels right to me. For data protection professionals, this changes the brief. We should not only ask whether data use is lawful. We should ask whether the system itself leaves the public sector stronger, weaker, freer, or more stuck. That is a better privacy conversation. More honest, more strategic, and much closer to the real world. Interested to hear how others are handling supplier dependency in privacy, procurement, or governance work. #DataProtection #DigitalSovereignty #Privacy #Data #DataSharing
-
James Robson shared thisPrivacy notices will not fix vendor lock-in. A lot of UK privacy work still focuses on the visible layer: notices, retention periods, DPIAs, data-sharing clauses. Necessary work. But it misses something bigger. If the systems underneath public services, healthcare, education, and government are heavily dependent on a small number of overseas providers, privacy stops being just a compliance exercise. It becomes a question of control, resilience, and public trust. That is why the digital sovereignty debate matters. The point is not to retreat from innovation, or pretend every global supplier is a problem. The point is to be honest about dependency. When core infrastructure is hard to inspect, hard to exit, and hard to challenge, risk builds quietly. Procurement teams feel it. Security teams feel it. Privacy teams should be talking about it far more than we do. The attached report makes a strong case that the UK has looked past this for too long. It links digital dependence to security exposure, weak bargaining power, poorer public value, and less room for democratic choice. It also argues that open standards, open source, and shared digital infrastructure offer a more durable path. That feels right to me. For data protection professionals, this changes the brief. We should not only ask whether data use is lawful. We should ask whether the system itself leaves the public sector stronger, weaker, freer, or more stuck. That is a better privacy conversation. More honest, more strategic, and much closer to the real world. Interested to hear how others are handling supplier dependency in privacy, procurement, or governance work. #DataProtection #DigitalSovereignty #Privacy #Data #DataSharing
-
James Robson liked thisExciting - only 84 sleeps now 😜James Robson liked thisIf you were curious about the rest of the IPDLN Conference 2026 workshop programme... Here is your third and final round of sessions 🔎 Joining the line-up on 13 July: 🔸 Integrating Health Data: Foundations and Frontiers in Record Linkage by Hye-Chung Kum, Toan Ong, Abel Kho, Shaun Grannis, Inessa Cohen 🔸 Data Governance Policy Frameworks and How You Can Use Them Today! by Jacob Pasner and Francis Mitrou 🔸 The Past, Present, and Future of Population Data Science by Emma Gordon, Andy Boyd, Kim McGrail, Scott Emerson 🔸 Assessing and Improving the Inference Quality of Linked Datasets by Utrecht University Register now to secure your place! Learn more and register: https://edu.nl/cdmrx #IPDLN2026 #DataLinkage #PopulationData #DataScience #Rotterdam #AcademicConference
-
James Robson liked thisI think it's great that LabourList is launching a supporter scheme. There was research last week that showed that LabourList is the second most-read morning email across MPs of all parties – it's so clearly a vital link between Westminster and membersJames Robson liked this🤔 Still thinking about becoming a friend of LabourList? 🔴 Here's one of our former reporters, Conor Pope on why you should join us 🫂 ❤️ Sign up here 👉️ https://lnkd.in/eY-3arpN
-
-
James Robson reacted on thisJames Robson reacted on thisThe Hidden Risk of AI on Enterprise Data The world's enterprise systems don't run on unstructured data. They run on structured data – billing systems, transaction logs, inventory records, customer accounts. And that's exactly where AI poses the greatest risk. In this first conversation with Michael Cation, founder and chairman of FractalWeb, we explore what it actually means to deploy AI safely against the systems your business depends on – and why getting it wrong isn't just a technical problem. It's an existential one. Most AI developments to date have focused on unstructured data like documents, text, images. But enterprise applications are built on relational databases, and giving an AI agent write access to a live system of record is a fundamentally different kind of risk. Corrupting a document is recoverable. Corrupting your global inventory management system is not. The approach FractalWeb has developed over 40 years addresses this directly. Rather than applying AI to live systems, they create a digital twin – a complete, synchronized copy of the system of record where AI experimentation can happen safely at scale. Each instance operates on a manageable slice of data, making it auditable and human-readable before any changes are promoted back to production. This is the first in a three-part series. Next, we go deeper into the architecture behind this approach — and explore how speed and safety can reinforce each other rather than trade off.
-
James Robson liked thisJames Robson liked thisLooking forward to the first'northern' Privacy Space today. Lovely drive over the peak district this morning.
-
James Robson liked thisJames Robson liked thisThe Whitaker Solutions Team all in one place, possibly for the first time ever! Zoë Freeman Silva Arapi, LL.M.
-
James Robson liked thisJames Robson liked thisI’ve agreed to be a judge for this year’s AbilityNet Tech4Good Awards. Thank you Robin Christopherson MBE for the invite. When people say “tech for good”… what has actually changed? What is the outcome? Something works differently. Someone is better off. A system behaves in a new way. If you’re building something like that, or you’ve seen a team doing it properly, put it forward. Some of the most meaningful work doesn’t shout the loudest. I’ll also be speaking to some of the finalists and winners on BASELINE to understand what actually got built, what had to be figured out, and where it holds up in practice. Deadline is 11 May! Mark Walker Gavin Ray. Stephen Warwick Gene Allmark-Kent David E. Rodney Foreman Eric Herzog Jack Perschke Bob De Caux Rachel Astall Sophie Costin Amy Low CPACC FRSA Daniel Sawko Dr John Fletcher Justin Keeble William Excell https://lnkd.in/ePX2yAkU
Experience & Education
-
TBC
View James’s full experience
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Licenses & Certifications
-
-
-
-
-
Certified Information Privacy Professional/Europe (CIPP/E)
IAPP - International Association of Privacy Professionals
Issued -
Certified Information Privacy Technologist (CIPT)
IAPP - International Association of Privacy Professionals
Issued -
-
-
-
Volunteer Experience
Courses
-
Coaching For Success - Company Course
-
-
Fire Warden Training
-
-
Recruiting From a HR Perspective - Company Course
-
-
St Johns Ambulance - First Aid Course
-
Projects
-
Short Films
-
See projectAn internet drama series based in Cardiff. This was a passion project made for no money about two weeks in the life of a Welsh Asian Pizza-Delivery driver.
It took two and a half years of work in which I wrote, produced and directed all 15 episodes and released over the summer of 2010 on the Media Wales website.
Recommendations received
18 people have recommended James
Join now to viewExplore more posts
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More