The Speech Police Are Worse Than We Thought

Police officers at a May Day rally in London, England, May 1, 2003. (Russell Boyce/Reuters)

The rise in speech-related offenses in the U.K. can be explained more by new laws than by any increase in objectively harmful conduct.

Sign in here to read more.

The rise in speech-related offenses in the U.K. can be explained more by new laws than by any increase in objectively harmful conduct.

I n 2025, the U.K. Times reported that police in England and Wales arrest 33 people per day — more than 12,100 per year — for speech offenses under Section 127 of the Communications Act (2003) or Section 1 of the Malicious Communications Act (1988). That alarming figure made headlines even in American publications, and it is regularly cited in discussions about England’s deterioration. In fact, Elon Musk himself has described the statistic as “insane.”

But that claim is severely misleading because it is limited to just two offense classifications and therefore downplays the abysmal state of free expression across the pond. A new study by an academic at Cardiff University in Wales analyzed police data across a much larger range of speech categories and distinguished between what it calls “core” and “non-core” crimes.


There are 1,192,200 recorded offenses across all these categories in England and Wales for the year ending June 2025, specifically 866,691 “core” crimes and 325,509 “non-core” crimes.

The former classification consists of legal offenses “where speech is not merely incidental, but comprises a substantive component of the offence relevant to the freedom of expression debate.” In other words, “core” crimes are those involving subjective thresholds for conduct that may be primarily expressive in nature and do not necessarily result in objective physical harm, specifically offenses coded as causing or committing (1) “Public fear, alarm or distress,” (2) “Racially or religiously aggravated public fear, alarm or distress,” (3) “Malicious communication,” (4) “Other offences against the State or public order,” (5) “Harassment,” or (6) “Racially or religiously aggravated harassment.” The study explains that police forces do not disaggregate physical conduct from speech when documenting harassment, thereby blurring how many charges involved speech alone. Of course, some harassment cases have been centered on politically tinged comments. For instance, comedian Graham Linehan was charged with harassment in April 2025 because he had made a series of social media posts about an LGBTQ+ activist, although he was eventually cleared. (However, he was convicted of a separate charge because he grabbed and threw the activist’s phone in the course of an altercation.)

By contrast, the “non-core” crimes are those that may be described as legitimately prohibited under expansive free-expression protections. In the study, “non-core” offenses are (1) “Obscene publications” and acts involving particular uses or types of pornography, (2) “Threat etc. to commit criminal damage,” (3) “Controlling and coercive behaviour,” (4) “Stalking,” (5) “Perjury,” and (6) “Threats to kill.” Although consensual pornography specifically is legal, the obscenity offenses may involve content featuring underage children or the illegal distribution of pornographic materials, among other things. Additionally, stalking may consist of speech, although it is generally understood as a behavioral pattern. Finally, perjury and death threats are condemned even by the most fervent supporters of the right to free speech. Put simply, these “non-core” categories are offenses that free-expression activists argue should not be protected, nor is there a particularly active public debate about what thresholds should be applied when evaluating these crimes.




We can argue whether the “core” and “non-core” classifications are neat enough, but in any event, the math works out to arresting (or, at least, documenting a charge against) more than 3,200 people each day. The total recorded crime amounts to 6,591,958 — meaning that 18 percent of all recorded crimes were broadly related to speech. More plainly, speech crimes amount to nearly one-fifth of all recorded crimes. Assuming that all the “non-core” offenses represent cases where legitimately criminalized speech was appropriately punished, then 13 percent of all recorded offenses potentially involve expressive conduct. (A minor qualification: Of the 13 percent, some may include instances of primarily physical harassment.)


It will not shock anyone that the conditions have worsened over time. The study found that “core” crimes grew by 442 percent from 2002–03 to 2024–25. Although the sheer amount of recorded “core” crimes declined after 2022, the trend reflects a difference in data collection. The Home Office adopted the “Principal Crime Rule” in 2023, a guideline for police to record the most serious offense in cases where multiple victim-based crimes are reported simultaneously against the same person; as a result, only a stalking count would be recorded when a person alleges both stalking and malicious communications. Therefore, the modest reduction in recorded speech offenses since 2022 cannot be interpreted as proof that these supposed crimes occurred at lesser rates, nor that prosecutions fell.

What should we conclude after reading this study? The overall rise in speech-related offenses is not evidence that people are increasingly committing objectively harmful conduct. Instead, the growth can be explained in part by new laws — like the Communications Act (2003) and the Equality Act (2010) — that directly or indirectly reduce the scope of permissible speech, thereby expanding the justifications to arrest someone for expression.


With that in mind, we should refrain from entirely blaming the police forces themselves for the current Orwellianism because, in some instances, they are doing what the legislation compels them to do. Still, policemen should not be entirely exculpated. Officers occasionally inflate the textual meaning of legislation (which is undoubtedly imprecise and criminalizes “grossly offensive” electronic communications) to pursue individuals who share controversial opinions on the internet because those individuals are easier targets than violent thugs. (Would you rather arrest the guy who posted an insensitive meme, or the guy wielding a knife?) But most important, these findings are unpleasantly sobering: We who were troubled by the fact that 12,000 people were arrested per year should now sound the alarm bells and declare a state of emergency.

Exit mobile version