The Human Filter: How UK Newsrooms Spot AI Pitches and What to Do Instead

AI has become a familiar presence in UK comms teams. It can summarise research, draft first versions, transcribe interviews and help sort a sprawling media list. Used well, it saves time. Used lazily, it produces the kind of media pitch that journalists now recognise within seconds: polished, plausible and strangely empty.

The backlash is not really about technology. It is about effort. Newsrooms are under pressure, inboxes are crowded and trust is finite. A pitch that feels assembled rather than considered wastes a journalist’s time and damages the sender’s reputation. The practical question for UK comms teams is therefore not whether to use AI, but where to draw the human filter.

The tells UK journalists are watching for

Journalists are not usually running forensic checks on every email. They do not need to. Generic AI-assisted outreach tends to reveal itself through patterns. These are some of the most common tells.

• The over-familiar opener. “I hope this finds you well” followed by a paragraph that could be sent to any journalist in the country is not personalisation. It is decoration.
• The missing local angle. A pitch about a national trend with no named town, institution, employer or reader consequence is difficult to place. Regional newsrooms in particular need to know why the story matters on their patch.
• The frictionless claim. Vague assertions such as “research shows” or “industry leaders agree” without a source, date or named expert are a warning sign. Journalists need verifiable material, not atmospheric confidence.
• The over-polished sentence. AI-assisted copy often has a uniform rhythm: every sentence the same length, every paragraph neatly balanced, every adjective safely chosen. It can read as if no human has decided what matters most.
• The hallucinated detail. A wrong title, a misnamed publication, an invented statistic or a quote attributed to the wrong person is more than an embarrassment. It signals that the sender has not checked the basics.
• The relentless follow-up. An automated sequence that asks for a response three times in a week, without acknowledging the newsroom’s workload, feels like pressure rather than a relationship.
• The absent human. If the pitch does not make clear who is accountable, who can speak and why that person is credible, the journalist has no reason to engage.

None of these tells is unique to AI. Bad pitching has always existed. What has changed is volume and speed. AI makes it easier to produce more outreach with less thought, and journalists have adjusted accordingly.

A human-first pitching checklist

A human-first pitch is not simply a document with fewer automation errors. It is a pitch where a person has made decisions: about the story, the audience, the evidence and the ask.

Before sending, apply a short checklist.

1. Is there a real story? Write the first sentence as if it were the headline. If it sounds like a corporate announcement rather than news, rethink the angle.
2. Does it belong to this journalist? Check recent bylines, not just the outlet. A business editor, a local government reporter and a features writer need different pitches.
3. Is the local relevance explicit? Name the place, the institution or the community affected. Do not leave the journalist to infer it.
4. Is the evidence checkable? Provide a source, a date, a named expert or a link to the underlying material. Avoid unattributed claims.
5. Is there a named human? Say who is available for interview, what they know and when they can speak.
6. Is the ask proportionate? A short email with a clear subject line is easier to assess than a long attachment and a request for a call.
7. Would you send this to a colleague? If the answer is no, it is not ready for a journalist.

This checklist is deliberately basic. Most media-relations failures are not sophisticated. They are the result of speed outrunning judgment.

Where AI earns its place

There is a legitimate role for AI in a comms team, provided it sits behind the human filter rather than in front of it.

AI can help with research: scanning public data, summarising reports, mapping coverage or identifying themes across a large set of articles. It can assist with monitoring, transcription and the early sorting of media lists. It can produce a rough first draft for internal discussion, or help a team stress-test whether a message is clear. It can also support data analysis and the production of background notes.

What it should not do is replace the judgment that makes a pitch worth reading. The decision about which journalist to approach, which angle matters, which evidence is credible and which spokesperson can carry the story is human work. AI can generate options. It cannot take responsibility for the relationship.

A useful internal rule is simple: use AI to widen the drafting and research process, but keep the final pitch, the local detail and the journalist relationship under named human ownership. That protects quality and makes accountability clear.

Authenticity as a measurable asset

Authenticity is often treated as a soft virtue. In media relations, it is becoming a practical asset. Journalists remember who sends useful, accurate and relevant pitches. They also remember who wastes their time. Over time, that memory shapes open rates, response rates and the willingness to take a call.

UK comms teams can measure this more deliberately. Track response rates by journalist and outlet. Note how often pitches lead to a conversation rather than a rejection. Ask spokespeople whether interviews felt well prepared. Monitor whether journalists return to the team for comment. These signals are not as tidy as a coverage count, but they indicate the health of the relationship.

The broader lesson is that AI has raised the premium on human judgment. When generic outreach is cheap to produce, relevance becomes more valuable. When polished copy is everywhere, specificity stands out. The brands that pitch well in 2026 will not be the ones that avoid AI. They will be the ones that use it without outsourcing the parts of media relations that depend on trust.

The human filter is not a restriction on efficiency. It is the quality control that makes efficiency worth having.