In defence of the Freedom of Information Act

What do these things have in common? The MPs’ expenses scandal; Prince Charles’s secret lobbying of ministers through the “black spider” memos; almost 20,000 complaints of sexual assault, abuse and harassment across mental health trusts in England since 2019; and, most recently, official figures showing that the government’s £5 billion pension tax raid will leave almost three million workers poorer in retirement.

Yes, again and again, the Freedom of Information Act 2000 (FOIA) has enabled journalists, campaigners and researchers to expose matters of public importance that an increasingly distant officialdom would rather keep out of sight.

Yet a coordinated effort to remake the Freedom of Information regime now appears to be underway across government and the Information Commissioner’s Office (ICO), with newly disclosed documents revealing proposals that would curtail precisely the kind of document-led scrutiny that FOIA was designed to facilitate.

Introduced under Tony Blair’s New Labour government, FOIA emerged from the buoyant post-Cold War conviction that the West — whose inhabitants were rapidly becoming familiar with the strange phrase “World Wide Web” — had defeated the closed societies of the second world principally because it was more open, transparent and accountable.

That self-flattering belief found its most complacent expression in Francis Fukuyama’s “End of History” thesis, in which liberal democracy marked “the end point of mankind’s ideological evolution”, and all that remained was for authoritarian societies to abandon their “pretensions of representing different and higher forms of society”.

It was in this self-satisfied milieu that Blair, then Prime Minister-in-waiting, declared in a 1996 speech to the Campaign for Freedom of Information that a Labour government would end the “obsessive and unnecessary secrecy which surrounds government activity”.

The Act that eventually emerged was less radical than the rhetoric that preceded it. Implementation was delayed for five years, and the idea that it would extend fully into Cabinet government was quietly abandoned. Nevertheless, section 1 still contained what one academic has described as a quietly revolutionary principle: “any person” making a request to a public authority — whether based in the UK or overseas — is entitled to be told whether the information is held and, if so, to have it communicated to them.

Yet disclosures released by the ICO under, ironically enough, FOIA reveal senior-level proposals that would significantly narrow that right. Notes from two meetings in March 2026 show ICO officials discussing FOI-related legislative questions with the Cabinet Office and other government officials.

An initial sign that reform was under consideration came on 18 March, when the Financial Times reported that ministers were exploring a reduction in the statutory cost ceiling under section 12 FOIA — in effect, lowering the point at which a public authority can refuse a request on the grounds that it would take too much time and resource to answer.

What the newly disclosed material suggests, however, is that other, similarly far-reaching ideas were also being discussed behind the scenes. Most revealing is a 6 March email to Warren Seddon, the ICO’s Director of FOI and Transparency, setting out an “initial suggested suite of provisions” aimed at “reducing the impact of requests… on public authorities and the ICO”, including requests that “have the effect of swamping [public authorities] and/or the ICO through their sheer number” or “contain significant superfluous … information/demands”.

Considered purely from an administrative perspective, the logic is understandable. In 2025, monitored bodies received 94,526 FOI requests, a 14 per cent increase on 2024 — and the highest annual total since monitoring began in 2005 — prompting a great deal of hand-wringing about the ease with which AI-assisted requests can now be generated, as well as concerns that some may be unduly complex or based on hallucinated premises.

Looked at another way, though, might those figures not hint at something altogether healthier: a public increasingly willing to harness digital technologies to make fuller use of a right Parliament gave them to probe for themselves how public authorities actually make decisions? It’s certainly true that 34,807 of those requests were withheld in full or in part. But for just over a quarter that was simply because they exceeded the cost limit. Only 940 — a whopping 2.7 per cent — were withheld as vexatious.

But of course this is the sunnier, practically Rousseau-esque reading of the figures. Whether the ICO sees them in quite the same light is, on the evidence of the disclosed proposals, rather less clear.

One of the simpler proposals floated in the disclosed “suite of provisions” is that individuals be limited to five requests per public authority per year, where currently there is no numerical cap. You don’t have to be a hard-bitten investigative journalist to spot the problem: FOI research is often iterative. A requester may begin with a broad scoping request, receive a section 12 refusal, narrow the request, receive partial disclosure, and then submit further targeted requests arising from the material itself. Five requests will therefore rarely mean five investigations; it may not even cover two.

Equally troubling is the proposal that “individuals” themselves could be declared “vexatious”. Existing ICO guidance on section 14 FOIA — the provision dealing with vexatious requests — makes clear that public authorities may consider the wider circumstances surrounding a request, including repeated requests, disproportionate burdens on staff or aggressive correspondence. But that is still a judgment about the request, viewed in context, not a standing status attached to the requester.

Even assuming this power came with safeguards, how long would such a label follow someone once the circumstances that gave rise to it had passed? People may behave obsessively, aggressively or irrationally for various temporary or situational reasons. A parent repeatedly submitting hostile requests to an NHS trust following the death of a child, for example, may generate difficult correspondence. Yet it is far from clear that such conduct should harden into a lasting form of bureaucratic stigma that shadows future requests.

What is so startling about this suggestion is that, in June 2025, the Commissioner criticised the Attorney General’s Office for failing to handle a journalist’s FOI request in an “applicant-blind” manner. Internal correspondence showed officials discussing how to “maximise the delay”, disclose information in the “most limited way possible”, and avoid creating something “easily quotable in a newspaper”. The ICO warned that the AGO’s use of section 14 risked giving the impression that it was being deployed as “an easier way to shut down awkward requests”.

“Awkward requests.” In some ways, this phrase goes to the heart of the tension here. Less than a year after rebuking the AGO, the ICO is proposing something uncomfortably close to the same approach, only now framed as a response to administrative burden. But at what point does “awkward” in the sense of burdensome slip back into awkward in the sense of politically inconvenient?

Perhaps the most constitutionally remarkable proposal is the suggestion that FOIA rights be restricted to UK residents and certain foreign nationals with voting rights. Whereas section 1 FOIA currently gives the right to request information to “any person”, the disclosed email canvasses turning that into a right exercised only by a narrower class of UK-based civic actors. The implications for overseas academics, journalists and NGOs conducting research into UK public authorities could be profound, potentially even collapsing research programmes that depend on sustained access to official information.

The cumulative effect of these proposals would be grim for investigative research. But the disclosed documents contain something still more consequential.

Buried in the small print is a proposal that section 14 FOIA could be “expanded” to “allow refusal of FOI requests in respect of which there is little public interest in its disclosure”. Admittedly, this is a slightly gnomic proposal. But it appears to point towards a possible recasting of FOIA from a right of public access, subject to defined limits, into something closer to official value-screening, in which officials increasingly decide which inquiries are sufficiently important to justify answering at all.

At present, public authorities have several ways to refuse requests, but two are especially relevant here. First, carefully delimited statutory “qualified exemptions” may be used where information falls within protected categories or where disclosure would cause specific forms of prejudice, such as inhibiting policy discussions, revealing legally privileged material or damaging commercial interests. Having first brought themselves within one of the recognised categories, public authorities must then ask whether the public interest nevertheless favours disclosure over maintaining the exemption, against the background of what ICO guidance describes as a general public interest in transparency. Refusal, in other words, must be tethered to one of the exemption categories.

Then there’s section 14 FOIA, which operates rather differently. Strictly speaking, it is not an exemption at all: it allows public authorities to reject vexatious requests because of the nature of the request, rather than the sensitivity of the information sought. Under Dransfield, the leading case on vexatious FOI requests, vexatiousness means something like a manifestly unjustified, inappropriate or improper use of FOIA. When deciding whether that threshold is met, public authorities may consider four broad themes: the burden on the authority; the requester’s motive; the value or serious purpose of the request; and any harassment or distress caused to staff.

Public interest therefore already has a place within section 14 analysis through the “value or serious purpose” limb. But unlike the qualified-exemption route, that assessment is not tied to fixed statutory categories and permits broader evaluative judgments about seriousness, legitimacy and public value. At present, however, it remains only one element within the wider, holistic vexatiousness assessment, alongside burden, motive and harassment.

That such judgments require careful handling, and should not be isolated from the wider assessment, is reflected in the ICO’s own guidance, which treats section 14 as a “high hurdle” precisely because of the need to preserve the Act’s underlying presumption in favour of transparency. Public authorities must consider “all the circumstances” before concluding that a request is disproportionate or unjustified. On “value or serious purpose”, the guidance warns authorities not to jump too quickly to conclusions about a request lacking value simply because its wider significance is not immediately obvious. Most requests, the ICO notes, will have some value and therefore some “reasonable foundation”, while requests prompted by private grievances may nevertheless overlap with wider public interests in transparency, accountability and institutional scrutiny.

On that basis, it’s difficult to see what practical work the proposed expansion would perform unless it made that public-interest judgment easier to deploy as a refusal ground. Otherwise, why expand section 14 at all? The natural implication is that the ICO is at least contemplating moving a consideration that currently operates as one factor in a wider balancing exercise closer to the centre of the refusal decision itself.

Is it too cynical to wonder whether “little public interest” is a more vague, subjective, pliable concept than burden, harassment or abusive motive, and therefore more readily deployed against those pesky “awkward” requests whose significance is not immediately obvious? Apparently not, since the email itself appears to respond to something very close to that danger. Immediately after floating the expansion of section 14, the author notes drafting difficulties around any objective test for “little public interest”, including whether such a test would have to be weighed against the likely impact on the public authority, before adding that “there are concerns about opening this up as it could be prone to abuse”.

Consider, for instance, an academic seeking internal correspondence surrounding a controversial redundancy process. At first glance, the request might appear to concern little more than a private employment grievance. Yet the same material could shed light on how a university conducts disciplinary negotiations, uses non-disclosure agreements (NDAs) or applies internal policies more broadly.

Under the current regime, that request would not automatically fail under section 14 simply because it appeared grievance-driven or unimportant at first glance, since requests prompted by personal circumstances may nevertheless overlap with a wider public interest, and authorities are not supposed to assume that a request lacks value or serious purpose simply because that value is not immediately self-evident.

FOI remains one of the few ways to get beneath the polished, PR-conscious, customer-facing language of public authorities

But an expanded “little public interest” ground could strengthen what is currently one consideration among several into something much closer to an independent basis for refusal, allowing public authorities to conclude more readily that disclosure lacks sufficient public value simply because, in their view, the request amounted to little more than a disgruntled employee pursuing a grievance — irrespective of whether the material might later prove relevant to a journalist examining NDA culture, governance failures or speech-related disciplinary practices.

Other than disclosure in costly legal proceedings, FOI remains one of the few ways to get beneath the polished, PR-conscious, customer-facing language of public authorities and examine how the sausage is actually made. But if public authorities are allowed to decide that a request is insufficiently in the public interest, requesters are limited to a handful of annual requests, the cost threshold for refusing requests is lowered still further, and overseas academics, journalists and campaigners with serious professional reasons to scrutinise UK public bodies are excluded altogether, the scope for complex investigations all but vanishes.

Not that this is FOIA’s first encounter with retrenchment. Ever since its practical commencement in 2005, successive governments have floated various schemes for restricting its scope: fees for requests, restrictions on heavy users, limits on institutional coverage, changes to appeal mechanisms and even removing Parliament from scope. In that sense, and notwithstanding their sweeping ambition, the ICO’s latest proposals are simply another link in the same chain.

Exactly thirty years on from Tony Blair’s speech to the Campaign for Freedom of Information, and his promise that freedom of information would help dispel the “disaffection” and “disillusion” engendered by secrecy, we appear to be tracing a wider historical arc, of which the attenuation of FOIA is but one manifestation.

No longer locked in an ideological battle with a rival social system making eyes at its citizens from behind the Iron Curtain, liberal democracies like the UK have nonetheless found another, altogether more introspective kind of war to fight. Buffeted by global financial pressures, and increasingly preoccupied with the impossible administrative demands of distributing dwindling resources across large populations, they show a marked tendency to treat the delicate constitutional attributes capable of pre-empting “disaffection” and “disillusion” as unaffordable luxuries.

And in the process, their slow transformation into something quite different is justified and managed not through the theatrics of spectacular authoritarianism but in the well-meaning, inoffensively bland technocratic idiom of cost containment and efficiency savings.