Britain Ends the Policing of Legal Speech Online — But the Records May Linger

On 31 March 2026, Home Secretary Shabana Mahmood announced that police in England and Wales will stop recording non-crime hate incidents (NCHIs) — the controversial practice that allowed officers to log and investigate social media posts and verbal exchanges that broke no law [1][2]. "Under these reforms, forces will no longer be policing perfectly legal tweets," Mahmood said. "Instead, they will be doing what they do best: patrolling our streets, catching criminals and keeping communities safe" [2].

The decision follows a joint review by the National Police Chiefs' Council (NPCC) and the College of Policing, which found that unclear guidance had led to officers attending people's homes over "insults and routine arguments" and that the system was "no longer fit for purpose" [3][4]. A government amendment to the Crime and Policing Bill has already revoked the statutory code of practice that underpinned NCHIs, replacing it with a narrower national standard tied to core police duties: preventing crime, safeguarding individuals, and maintaining public order [5].

The announcement closes one chapter of a years-long argument over where policing ends and free expression begins. It opens another over what happens to the tens of thousands of records already created, and whether the categories of harmful-but-legal content that NCHIs were designed to flag will now go entirely unmonitored.

The Scale of the Problem

Two parallel enforcement mechanisms have drawn criticism. The first is NCHIs themselves: over 133,000 were recorded by 43 police forces in England and Wales since 2014 [6]. In the most recent reporting year (2024–25), 9,305 incidents were logged across 34 forces [1]. While the annual total has been declining — from roughly 14,270 in 2020–21 to 9,305 in 2024–25 — the cumulative volume means a substantial number of individuals have records attached to their names for conduct that was never criminal [3].

Source: NPCC / College of Policing Review

Data as of Mar 31, 2026CSV

The second mechanism is outright arrest. Freedom of Information data compiled by Big Brother Watch and reported by The Times showed that in 2023, UK police made 12,183 arrests under Section 127 of the Communications Act 2003 and Section 1 of the Malicious Communications Act 1988 [7][8]. That figure represents roughly 33 arrests per day. It had more than doubled since 2017, and represented a 58% increase over pre-pandemic levels [7]. Yet fewer than 10% of those arrested — some 1,119 — were convicted and sentenced, a disparity that critics describe as evidence of systematic overreach [7][9].

Source: Big Brother Watch / FOI Requests

Data as of Nov 17, 2025CSV

By 2024, the annual arrest figure under these statutes fell to roughly 9,950, according to FOI responses from 39 police forces [8]. The decline coincided with growing public and political scrutiny, including the Metropolitan Police's own decision in October 2025 to stop investigating NCHIs unilaterally, five months before the Home Office made the policy nationwide [10].

The Legal Architecture That Enabled Overreach

Multiple statutes gave police broad latitude over online speech. Section 127 of the Communications Act 2003 criminalises messages that are "grossly offensive" or sent for the purpose of causing "annoyance, inconvenience or needless anxiety" — language that civil liberties organisations have long argued is too vague to serve as the basis for criminal prosecution [9][11]. Section 1 of the Malicious Communications Act 1988 covers messages intended to cause "distress or anxiety." The Online Safety Act, which took partial effect in 2023, added further obligations; The Telegraph reported that 292 people had been charged under its provisions by February 2025 [9].

NCHIs sat outside the criminal framework entirely. They originated from Recommendation 15 of the 1999 Macpherson Report into the murder of Stephen Lawrence, which called for police to record racist incidents even when no crime was alleged, as an intelligence tool for identifying escalation patterns [6][12]. Over subsequent decades, the category expanded beyond race to cover any incident perceived by the complainant as motivated by hostility toward a protected characteristic — a "perception-based" standard that courts later found problematic [13].

The NPCC/College of Policing review identified several mechanisms of overreach. The national standard governing when police should record an NCHI had not been updated since 2011, well before social media became a dominant arena for public speech [3][4]. Individual forces applied the standard inconsistently: some recorded every complaint referred to them; others applied informal filters. Third-party reporting — complaints filed not by the person affected but by organisations or bystanders — contributed to volume, though the review stopped short of naming specific organisations [4].

The Court Rulings That Preceded the Policy Change

The legal challenge brought by Harry Miller, a former police officer who was visited at his workplace in 2019 after posting gender-critical tweets, proved decisive in exposing the system's flaws. In February 2020, the High Court ruled that Humberside Police had unlawfully interfered with Miller's right to free expression under Article 10 of the European Convention on Human Rights by recording his tweets as an NCHI and sending an officer to warn him [13][14].

In December 2021, the Court of Appeal went further, ruling that the College of Policing's 2014 guidance on NCHI recording itself was unlawful. The court held that the guidance "sanctioned or positively approved or encouraged unlawful conduct" and ordered additional safeguards to ensure that "the incursion into freedom of expression is no more than is strictly necessary" [13][14]. The College of Policing subsequently revised its guidance, but pressure for outright abolition continued to build from both the courts and Parliament.

International Comparison

The UK is not alone in grappling with the policing of online speech, but its enforcement volume stands out among comparable democracies.

Germany's Network Enforcement Act (NetzDG), enacted in 2017, placed obligations on platforms rather than individual users — requiring social media companies to remove illegal content or face fines of up to €50 million [15]. German prosecutors do pursue individuals for online speech, particularly Holocaust denial and incitement to hatred, but enforcement is channelled through the judicial system rather than through police recording of non-criminal incidents [15][16].

France follows a broadly similar model under the EU's Digital Services Act. Research by the Future of Free Speech project found that across Germany, France, and Sweden, between 87.5% and 99.7% of social media comments deleted under platform compliance rules were classified as legally permissible speech — raising concerns about over-removal, but through private-sector moderation rather than police action [17].

Australia has pursued social media regulation through the eSafety Commissioner, an independent statutory body with powers to issue takedown notices for certain categories of content. The commissioner's remit focuses on cyberbullying, image-based abuse, and illegal content rather than on broad categories of offensive speech [9].

Canada's approach has centred on human rights tribunals and, more recently, the proposed Online Harms Act (Bill C-63), which would create a Digital Safety Commission. The bill has drawn criticism for potentially enabling complaints about lawful speech, echoing concerns raised about the UK's NCHI system [9].

What distinguishes the UK is the scale of direct police involvement with individual users — arrests, home visits, and formal recording of non-criminal conduct — rather than platform-level regulation. Freedom House's 2025 Freedom on the Net report downgraded the UK's score, citing the volume of arrests for online speech and noting that "some civil liberties groups expressed concern that the laws were being applied broadly and in some cases punished speech protected by international human rights standards" [9].

What Happens to Existing Records

The government's announcement does not automatically expunge the records already held. Over 133,000 NCHIs logged since 2014 remain on police databases [6]. For individuals who were recorded, the consequences can be lasting. Enhanced DBS (Disclosure and Barring Service) checks — required for work in education, healthcare, and other regulated sectors — draw on the full range of police-held information, including intelligence that never resulted in a charge or conviction [18].

Under current filtering rules, some cautions and spent convictions are automatically removed from DBS disclosures after a set period [19]. But NCHIs occupy a grey area: they are not convictions, yet they sit on police intelligence systems and can be disclosed at a chief officer's discretion if deemed relevant to the role being applied for [18]. The Home Office has not announced any programme to review or delete historical NCHI records, and the new national standard focuses on future recording rather than retrospective cleanup [5].

This gap has drawn criticism from multiple directions. The Free Speech Union has argued that individuals recorded for lawful speech should have those records deleted entirely [20]. Big Brother Watch, while welcoming the scrapping of NCHIs, cautioned that without an active review of existing records, the chilling effect on free expression would persist for years [8].

The Case for Monitoring Harmful-but-Legal Content

Not everyone views the end of NCHIs as a straightforward victory for civil liberties. Baroness Doreen Lawrence, whose son Stephen's murder in 1993 prompted the Macpherson Report that led to NCHI recording, told the House of Lords that the system served as an early-warning mechanism. "It depends on how you see non-crime hate and it depends on who's at the receiving end of that," she said. "Individuals who think they've got the right to walk around and talk about, especially, young black men in a certain way, what starts off as just verbal, it leads to violence" [12][21].

Hate-crime researchers have pointed to evidence that verbal abuse and online hostility can precede physical violence — a pattern documented in the context of the August 2024 riots, when misinformation spread on social media contributed to attacks on mosques and asylum seeker accommodation across England. More than 30 people were arrested for social media posts connected to the riots, and 1,840 people were arrested overall by July 2025 [22][23]. Some researchers argue that a system for tracking sub-criminal hostility could have flagged escalation risks before the violence began [12].

Domestic-abuse organisations have raised related concerns. Patterns of coercive control and targeted harassment often involve conduct that, taken in isolation, falls below the criminal threshold but that in aggregate constitutes sustained harm. Without a recording mechanism, those patterns become harder to demonstrate to courts [21].

The government's position is that the new national standard preserves the ability to record incidents where there is a "clear risk of harm" — the distinction being that the threshold is now tied to policing purpose rather than the perception of any complainant [4][5]. Whether that threshold proves workable in practice remains to be tested.

Who Commissioned the Review, and Why Now

Home Secretary Mahmood commissioned the urgent review of NCHI guidance from the NPCC and College of Policing shortly after taking office [3][4]. In September 2025, the NPCC and College of Policing wrote to the Policing Minister recommending the immediate abolition of NCHIs in their current form [3]. The government accepted all of their recommendations [4].

The timing raises questions. The Labour government faced sustained criticism over the summer 2024 riots for the volume of social media arrests, with figures including Elon Musk publicly accusing the UK of authoritarian speech policing [8][9]. The Conservatives, then in opposition, had already been calling for NCHIs to be scrapped [20]. Shadow Home Secretary Chris Philp dismissed the March 2026 announcement as "simply a rebrand of non-crime hate incidents with a more restrictive triage process," arguing that "reports are still logged, personal data still recorded, and disclosure rules are unchanged" [10][24].

Whether the reversal represents a genuine rights correction or political repositioning is a matter of interpretation. The review was conducted by policing bodies rather than an independent commission, and its terms of reference focused on operational effectiveness rather than civil liberties principles [4]. No explicit finding was made about political pressure on individual policing decisions, though the review did note that ambiguous central guidance — rather than ministerial direction — was the primary driver of inconsistent enforcement [3][4].

Accountability for Past Overreach

The Miller litigation remains the most prominent example of individual accountability. The High Court's 2020 finding that Humberside Police acted unlawfully established a precedent, and the Court of Appeal's 2021 ruling that the College of Policing guidance itself was unlawful went further [13][14]. But those rulings addressed systemic guidance rather than individual officer conduct.

No public reports indicate that formal misconduct proceedings have been initiated against officers who conducted NCHI investigations now deemed to constitute overreach. Civil claims are possible under the Human Rights Act 1998 for individuals who can demonstrate that their Article 10 rights were infringed, but such claims require resources and legal representation that most affected individuals lack [13].

The government has not announced any review of past cases, compensation scheme, or mechanism for individuals to challenge historical NCHI records. This stands in contrast to other jurisdictions where policy reversals have been accompanied by remedial measures — such as the expungement of cannabis convictions in several US states following legalisation [5].

What Changes — and What Doesn't

The practical effect of the announcement is that the statutory code of practice governing NCHIs will be revoked through the Crime and Policing Bill [5]. The police database used to record crimes will no longer be used to log hate-motivated incidents that fall below the criminal threshold. Where incidents are recorded under the new standard, the records will not use crime terminology such as "victim" and "suspect" [4].

But the underlying criminal statutes — Section 127 of the Communications Act, Section 1 of the Malicious Communications Act, and the Online Safety Act — remain in force [9][11]. Arrests for online speech that police deem criminal will continue. The question is whether the removal of the sub-criminal recording layer will meaningfully reduce the volume of police contact with individuals over their online expression, or whether, as Philp argues, the change amounts to relabelling rather than reform [10][24].

For the estimated 133,000-plus individuals whose non-criminal speech was logged on police databases over the past decade, the immediate answer is: nothing changes. Their records remain. The policing of legal speech in Britain may have formally ended, but its consequences have not.