Complaints about automated nuisance calls almost tripled in three months, the ICO says, and it blames a surge in AI-enabled marketing calls. Its own press release cites no AI-specific data anywhere to support that claim.
Estimated reading time: 5 minutes
In short. The Information Commissioner's Office says complaints about automated nuisance calls rose from 1,374 in July 2026 to 3,684 in September, "reaching its highest level since 2021" [[1]]. It frames that rise amid "growing concern about automated and AI-enabled marketing calls" [[1]], but nowhere in the release, or in the regulator's own separate guide to what a nuisance call is, does an AI-specific figure, category or source appear [[1]][[2]]. The release's only commitment to act, "a number of live investigations... under way" [[1]], carries no number, sector or date. Two documented mistakes: four stars.
Three monthly totals, published by the regulator itself, are hard to argue with: complaints about automated nuisance calls to the Information Commissioner's Office rose from 1,374 in July 2026 to 2,178 in August, "an increase of 58%", then to 3,684 in September, "a further 69%" on top of that [[1]]. The ICO's own headline calls it a "five-year high" [[1]]. Andy Curry, the ICO's Head of Investigations, says plainly: "The numbers we're seeing are likely to be just the tip of the iceberg" [[1]]. Nothing in that part of the release is in dispute here.
What sits next to those numbers is a different kind of claim. "The increase comes amid growing concern about automated and AI-enabled marketing calls, commonly known as robo calls, as well as other calls which can be used to trick people into joining WhatsApp groups, sharing personal information or engaging with fraudulent schemes," the release says [[1]]. Three precise, dated, sourced figures sit beside an undated, unsourced technology story, and the release moves between them as though they were the same kind of evidence.
The claims, tested
| The document's own words | What the cited evidence shows | Verdict |
|---|---|---|
| "The increase comes amid growing concern about automated and AI-enabled marketing calls" [[1]] | The release's complaint totals (1,374, 2,178, 3,684) are a single figure for "automated nuisance calls"; no number, study or source anywhere in the release breaks out how many involve AI, and the ICO's own separate guide defines only two categories of nuisance call, "live marketing calls" and "automated marketing calls", with no AI-specific type [[1]][[2]] | A measured, sourced rise in one category is used to support an unmeasured claim about a different, undefined one |
| "We are actively investigating the rise in automated calls" [[1]]; "We have a number of live investigations under way into nuisance and scam calling activity" [[1]] | No number of investigations, no named sector or company type, and no date is attached to either sentence; the release's only quantified enforcement record, 15 Monetary Penalty Notices totalling £2,370,000, spans April 2025 to date and is not linked to the July-to-September 2026 rise it sits beside [[1]] | The one sentence in the release promising a response to this specific spike is also the one sentence with no figure attached to it |
A category that does not exist yet
The ICO publishes its own definition of what it is counting. Its guide to nuisance calls, aimed at the public, states: "There are two types of nuisance calls: live marketing calls, which are calls from a real person, that you haven't consented to receive; and automated marketing calls, which are pre-recorded marketing messages that are played when you answer the phone" [[2]]. That is the whole taxonomy. It is also, presumably, the categorisation behind the three monthly totals the press release reports, since the release's own figures are headed "automated nuisance calls" and nothing else. Neither category mentions how a call was generated, let alone whether artificial intelligence was involved in writing, voicing or targeting it.
That gap matters because "AI-enabled" is doing real work in the release's framing, not decorative work. The sentence that introduces it groups it with "robo calls" and separately, "as well as", with calls used for WhatsApp and fraud scams, implying AI-enabled calls are a distinct, growing slice of the problem worth naming specifically. But a regulator that collects complaints under a two-category system with no AI field cannot, from that data alone, say whether AI involvement is rising, falling or flat within the tripling it has just reported. The claim may turn out to be correct. It is not, on this document's own evidence, a finding; it is a framing choice attached to a number that cannot support it.
A promise with no numbers in it
The release is precise everywhere it wants to be. Three monthly complaint figures, two percentage increases, one five-year comparison, 15 Monetary Penalty Notices, one combined total of £2,370,000, two named companies fined £550,000 between them. Then, describing what the ICO is actually doing about the rise it has just quantified: "We are actively investigating the rise in automated calls and are urging the public to report their experiences" [[1]], followed later by "We have a number of live investigations under way into nuisance and scam calling activity and has recently taken enforcement action against organisations responsible for unlawful marketing calls" [[1]]. No count attaches to "a number of". No sector, platform or calling method is named. No date is given for when an outcome might be expected.
The £2,370,000 figure sits in the same paragraph as if it answered the question, but it does not. It covers 15 penalty notices "since April 2025" [[1]], an eighteen-month span that happens to include the July-to-September spike the release describes, but the release never says how many of the 15, if any, relate to it. The two named examples, Home Improvement Marketing Ltd and Green Spark Energy, are both described as fined "last year" [[1]], meaning before the three-month rise being reported now. A reader is given a historical enforcement total standing in for an answer about a current one.
Credit where due
The ICO is not hiding the scale of the problem it is reporting. The three monthly figures are specific and, unlike the AI framing around them, fully sourced to the regulator's own complaint data. Andy Curry's quote is candid that the true number is likely higher than what gets reported, rather than taking the complaint total as the whole picture. The release also gives the public a working action, the online reporting tool, rather than only asking them to be vigilant. The gap here is not in what the regulator discloses about the scale of the problem; it is in what it asserts about the cause, and what it commits to about the response.
Verdict
Four stars, from two documented mistakes. The underlying data, a sharp and genuine rise in nuisance call complaints, is real and the ICO reports it straight. The weak points are the two things layered on top of that data: a technology-specific cause the regulator's own complaint categories cannot currently measure, and an enforcement promise with no number, scope or date attached to it.
The star score counts two documented mistakes. First, the release attributes a measured rise in "automated nuisance calls" complaints partly to "growing concern about automated and AI-enabled marketing calls" without citing any AI-specific figure, study or source, and the ICO's own separate guide to nuisance calls defines only two categories, live and automated, with no AI-specific type. Second, the release's only stated response to the rise, "a number of live investigations... under way", carries no count, sector or date, and the one enforcement total it does give, 15 Monetary Penalty Notices since April 2025, is not linked to the specific three-month rise it sits beside in the text. Two falls in the 1 to 3 band: four stars; the bands are on the ratings page. This piece makes no finding about the competence of Andy Curry or any other named individual; the criticism is of the department's published framing and commitments, not of any person's conduct. Checked directly against the ICO's press release and its separate public guide to nuisance calls, both fetched 10 October 2026.
Sources
- Complaints about automated nuisance calls hit five-year high, Information Commissioner's Office, 9 October 2026
- Nuisance calls, Information Commissioner's Office