What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
AI governance cannot be reduced to model performance or a list of ethical principles: it also has to account for the institutions, incentives and power relations that shape how systems are built and used. That was the central thread in Computer Weekly’s account of techUK’s eighth annual Digital Ethics Summit, held in December 2024, where speakers discussed how to make AI ethics practical, inclusive and relevant to real deployments.
What does it mean to treat AI as socio-technical?
A socio-technical view recognises that technology and social processes shape one another. AI systems are built from technical components, but decisions about data, design, procurement and deployment are made within organisations and societies. Once in use, those systems can affect people’s opportunities, public trust and the distribution of power.
Computer Weekly reported that public officials, industry figures and civil society groups at the summit considered the ethical challenges created by AI’s spread and what might lie ahead in 2025. Delegates called for AI to be governed as a socio-technical system, rather than treating ethics as a separate layer added to a finished product. The event account records discussion and individual views, not a formal summit consensus.
The article described a contemporaneous expectation that 2025 would be a “year of diffusion.” That was a forecast made in late 2024, not a verified account of what subsequently happened or a current prediction.
#1 Best Overall
Why are principles not enough?
Ethical frameworks can set direction, but speakers said that high-level principles often leave organisations uncertain about what responsible practice requires in a particular application. The challenge is to translate principles into decisions that can be explained, tested and governed.
Explainability depends on the use case
Leanne Allen, then KPMG’s UK head of AI, said established principles remained durable but difficult to apply. She described explaining generative AI outputs as “fundamentally difficult” and called for more nuance and guidance on what principles mean in practice. The point is not that explanation is impossible in every setting, but that a general promise of explainability does not by itself specify what an organisation should disclose or demonstrate for a particular system.
Audits and bias evaluations lack shared definitions
Melissa Heikkilä, then a senior reporter at MIT Technology Review, said that audit and bias-evaluation methods varied between companies and that a lack of meaningful transparency obstructed standardisation. As Computer Weekly quoted her: “I think no one can agree how to do a proper audit, or what these bias evaluations look like. It’s still very much in the Wild West, and companies each have their own definitions.” The report did not supply comparative audit results or a common measure of audit quality.
Standardisation must stay grounded in deployment
Alice Schoenauer Sebag, then a senior member of technical staff on Cohere’s AI safety team, argued that shared standards could support a common understanding and innovation. She pointed to the risk and reliability working group at MLCommons as work toward a shared taxonomy and benchmarking. At the same time, she stressed that assurance has to address the intended application: “I wouldn’t necessarily say that the [ethical] conversations are getting harder, I would say that they’re getting more concrete.” She said discussions with customers were becoming specific to “what it means for this application, this use case, to be deployed and to be responsible.”
Free tools Windows power users keep installed
One-click scans. No signup required.
That application-level focus matters especially when an organisation buys rather than builds an AI system. Allen said buyers often depend on supplier contracts for assurances about ethical standards, while lacking direct control over how the system was developed. She also warned that assurances remain imperfect and uncertainty persists. A contract can allocate responsibilities, but the account does not present it as a substitute for scrutiny of the system in its actual setting.
Who should shape AI systems and their benefits?
Participants called for people affected by AI to be involved early in a system’s lifecycle, not only consulted after its design or deployment. Jeni Tennison, founder of Connected By Data, described public deliberation and user-led research as ways to work with civil society and the public. Her invitation was: “Let’s together find the route that leads us to something that we all value.”
Rank #3
Inclusion includes language and geography
Schoenauer Sebag cautioned against defining safe AI from a Western, English-centric perspective, arguing that products deployed around the world should be safe in ways that make sense in different places. Heikkilä likewise connected language representation to power: in her assessment, concentration of data and compute among a small number of firms and countries could grow, making diverse languages and geographic representation important considerations. These were speakers’ concerns, not quantified findings about language coverage or measured forecasts of concentration.
Distribution within a country matters too
Andrew Pakes, then the Labour (Co-op) MP for Peterborough, argued that policy should consider inequalities within the UK as well as geopolitical competition. He contrasted Peterborough with nearby Cambridge to question whether AI’s gains would reach beyond established innovation centres. “We have two different lives that people live just by the postcode they live in – how do we deal with that challenge?” he asked. He warned that people need to feel change is being done with them rather than to them, or society could lose economic benefits and experience deeper division.
Hetan Shah, then chief executive of the British Academy, invoked austerity and the “Big Society” to warn that presenting reduced public services as inclusion could backfire. “If that’s where AI gets wrapped up, citizens won’t like it. Your agenda will end in failure if you don’t think about the citizens,” he said. The warning links public acceptance not just to the technology, but to what AI is being used to justify or replace.
What role should government play in public-sector AI?
In a panel on responsible AI diffusion through public services, speakers discussed co-design and early engagement alongside government’s ability to act as more than a regulator or customer. Alex Krasodomski, then director of Chatham House’s Digital Society Programme, argued that governments need the capacity and mandate to build public AI services. In his view, that capability would help them negotiate with suppliers on more equal terms, instead of depending entirely on large technology companies for systems their populations rely on.
The account also presents competing considerations around public capacity: infrastructure is expensive and requires scale, while reliance on a small number of providers can create dependence. Chloe MacEwen of Microsoft discussed UK cloud infrastructure and the possibility of third-party audits and assurance as markets. Linda Griffin of Mozilla characterised cloud concentration as a geopolitical issue. These are perspectives from speakers representing different organisations, not independent findings about the UK cloud market.
In that discussion, the article referred to Microsoft’s December 2023 announcement of a £2.5bn investment commitment for the following three years. It was contextual information reported as part of a speaker’s remarks, not a summit finding or a measure of AI ethics, safety or outcomes.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchMartin Tisné, then CEO and thematic envoy to the planned AI Action Summit in France in early 2025, advocated international collaboration. The event account placed that call in the context of the Bletchley Park AI Safety Summit in November 2023 and the AI Seoul Summit in May 2024; those dates describe the historical context as reported in December 2024.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Does the summit report show that open or closed AI is safer?
No. Computer Weekly’s account records debate rather than a conclusion that either open or closed approaches are inherently safe. Speakers discussed openness as a way to widen access and enable innovation, while also raising safety concerns. Griffin’s position was that both open and closed systems need guardrails. She argued for greater openness to help people understand systems used at scale in areas such as healthcare, including the data and training behind their decisions.
| Question | Why it matters in the open-versus-closed debate | What the summit account establishes |
|---|---|---|
| Access and scrutiny | How much information about a model is available, and who can inspect or adapt it? | The report records arguments that openness can widen access and innovation, and that understanding systems can support trust. It does not establish that access alone ensures effective scrutiny. |
| Control and dependence | Who sets the terms of deployment, and how dependent are users on a provider, API or cloud service? | These were identified as relevant governance questions; the account does not rank approaches on control or supplier dependence. |
| Safety and guardrails | What risks arise in a specific use, and what safeguards can address them? | Speakers argued that both open and closed systems need guardrails. No comparative safety outcome was reported. |
| Fit for people and context | Does the model and its governance fit the application, affected communities and languages involved? | The report supports context-specific assessment and inclusion, not a universal preference for either approach. |
The article also quoted a US National Telecommunications and Information Administration report as saying: “Current evidence is not sufficient to definitively determine either that restrictions on such open weight models are warranted, or that restrictions will never be appropriate in the future.” In Computer Weekly’s account, this underscores uncertainty rather than resolving the debate.
How should readers interpret the summit account?
This is a retrospective of one journalist’s account, published by Computer Weekly on December 17, 2024. It is not an official transcript, a record of formal consensus, or a current assessment of policy, markets or developments after the event. Speaker quotations and positions here are attributed as reported by Computer Weekly; the account provides no summit-specific quantitative finding on AI-driven inequality, trust, audit quality, language coverage or concentration.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
