Don’t scrap DSIT — refocus it

There is a good idea here, hiding inside a bad one

Jul 20, 2026

It was always going to be risky, changing Prime Ministers in silly season. The rumour mill is spinning even faster than normal. This weekend, there were reports that Burnham will scrap the UK’s Department for Science, Innovation and Technology; some suggested that this shows the new government is not on the side of Silicon Valley. It’s not clear yet if this argument fell half-baked from camp Burnham, or was briefed out by a civil servant. Whether true or not, it is a bad idea as stated. But like many bad ideas, it hides a good idea within it.

Let’s start with why dismantling and demoting the government’s digital capabilities would be foolish. Most obviously, it would cut the legs from under Burnham’s core devolution agenda, which is all about unleashing the innovative potential of places and communities across Britain.

It’s long been known that one of the key constraints on local innovation is the tangle of legacy outsourced IT systems that binds local government. This is why shared public digital capabilities, like the ones in DSIT, are important. Leaving every council to build its own tech stack — or buy one from a failed software marketplace — is about as empowering as handing out straitjackets.

It’s also clear that you can’t do genuine place-based decision-making without the capability to integrate data across silos. This again is precisely the kind of capability that sits in DSIT, and that can now be put more squarely in service of places and communities. Dismantling this capability would hobble one of the most enabling parts of Whitehall, just as new devo deals are negotiated.

There is a wider point here too. In the last few years, the frontier of public digital work has turned outward, starting a new era of digitally-enabled devolution. We described this recently, in a report from Kinship Works and Platformland, via a series of operating patterns — a way of thinking about devolved systems in a digital age. We did not invent these patterns; we were just codifying the leading work that is happening. Take India as an example: their public tech stack combines the power of digital with a decentralising instinct. The centre builds rails (e.g. for capabilities like identity, payments, and data-sharing) that states, cities, and startups innovate on top of, boosting GDP in the process. In the last five years, the best UK public digital work has embraced this same decentralising spirit, from GDS Local to i.AI’s work with local planning officers. Lose 18 months on an internal re-org of these teams, and we slow down all these efforts.

So a strong public digital capability is vital for devolution. But it is also vital for another of Burnham’s agendas — his recognition of the threat posed by Big Tech to democracy, epistemic security, and even national sovereignty. If Burnham is right to make these arguments, which he surely is, then restructuring the government’s digital capabilities right now would be like putting down our weapons. Instead, we need to follow through on another point Burnham allies have been making — we should see core digital capabilities as vital public infrastructure, no less important than roads and railways.

This takes us to a wider way in which Burnham’s agenda aligns with the frontier of public digital work, including work already happening in DSIT. As I wrote in a recent post, today’s leading public digital work recognises the need for market-shaping capabilities. No government that is serious about digital today — or indeed about democracy and national sovereignty — can shy away from shaping the digital markets in which the state and public life operates. This means taking a more activist stance: building public options to break the hold of monopolies and duopolies; mandating standards like interoperability to liberate institutions — especially local government — from lock-in to proprietary systems; and even making strategic acquisitions. It also means investing in legal and procurement expertise, to help exit legacy contracts and avoid getting stuck in new ones. Again, this is work that has already started in DSIT, which has been investing in its commercial capabilities. This work, too, should now be accelerated.

Where does all of this point us? It suggests that, while it would be daft to scrap DSIT, there is a case for refocusing the department’s digital efforts. Give these teams an explicit mandate to support a radical devolution agenda — not just empowering places and neighbourhoods, but also frontline public servants. Take a leaf out of India’s book, building the rails of a powerfully enabling public tech stack. Rather than slicing and dicing the department — moving parts to DBT, separating AI from GDS, and leaving digital work as an afterthought in DCMS — create a sharper and more powerful Department for Science, Digital and Local Innovation.

What this means, in short, is turning DSIT outward. That could include doubling the size of GDS Local, complemented with a more explicit mandate for GDS itself to support local innovators. Set up a sister organisation to GDS Local for the NHS, working hand-in-hand with innovators in NHS Trusts and other local NHS bodies — not least to support neighbourhood healthcare. Learn from MHCLG’s Local Digital Fund, which helped seed initiatives like LocalGov Drupal’s open-source web publishing for councils, and run a bigger, better version. Spin up more experiments of the kind i.AI are doing on planning — saving hours for local officers, freeing people to engage in differentiating and strategic local work.

Alongside this, a Burnham-appointed Secretary of State at the new Department could frame the government’s digital work explicitly as building a much-needed public tech stack. Recognise, finally, that digital capabilities are vital public infrastructure, borrowing from the playbook of pioneers like Audrey Tang in Taiwan — and work with the Treasury to make sure the work is funded in this way. A Burnham government seems set on ditching ID cards, but clearly that should not mean ditching the capability to verify identity in order to access public services — without relying on the big US tech platforms. To complement this, work could be accelerated and deepened on a full suite of public digital infrastructure.

What about DSIT’s wider portfolio? I have ignored broad swathes of work — from the science budget, funding for universities via UKRI and the research councils, and industry-facing work, such as that done by Innovate UK. This includes initiatives like the £100m Innovation Accelerators — 26 projects across Greater Manchester, the West Midlands and Glasgow, which Burnham called “a new, place-based model for innovation-led growth” — and the £500m Local Innovation Partnerships Fund, now scaling to regions across the UK. See, as an example of the potential here, Jamie Hardesty’s work and the North East AI plan. Here there is a case for moving the more industry-facing parts of DSIT’s work into a Ministry for Industry, working with Combined Authorities on place-led growth strategies.

If DSIT lost these aspects of its work, would that leave enough to justify a department and a place in Cabinet? Of course it would. Indeed digital work in today’s world is now so deeply political, and still so vulnerable to Whitehall inertia, that it is vital that this work is represented at the top table. Plus I haven’t touched on a whole wider agenda that could be embraced by a new, more radical refocused department. That could include supporting communities by fostering an ecosystem of public interest technology. And it could include emergency work to defend democracy — using AI and deliberative platforms to begin clawing our way back to a healthier digital public realm.

What this all adds up to, I would argue, is less of a dialling down of digital work, and more a change of flavour and emphasis. Root out any remaining tendencies towards centralisation and ‘Whitehall-knows-best’. Recognise the state’s role building public options, and shaping markets in the public interest. Give the work less of a flavour of the Tony Blair Institute’s tendency to solutionism, and put a bit less emphasis on personalisation, choice, and competition. Instead, emphasise such ideas as public value and purpose, and the power of places, communities, and frontline workers as innovators and discovery mechanisms.

All of which is to say: don’t hobble the work in restructures, or split AI and digital off to different ends of Whitehall. Instead, make the work more powerful by giving it a new vision and focus — and then turn it outwards, to support a more devolved and dynamic public sector.

For more in a similar spirit, see this post on Test & Learn methods in the NHS. Or this post on the wider question of how we can save bureaucracy from itself. You can follow my writing on Blue Sky, Medium, or Substack. If you’d like to work together on these topics, drop me a line at Kinship Works — we support more adaptive, contemporary, and human approaches to government.

Future State

Discussion about this post

Ready for more?

Future State

Don’t scrap DSIT — refocus it

There is a good idea here, hiding inside a bad one

Other writing mentioned in the post:

Discussion about this post

Ready for more?