The AI News Gatekeepers: Why ChatGPT Cites The Guardian 58% of the Time

New research reveals that AI tools are creating a winner-take-all publishing hierarchy—and most brands don't realize they're already locked out

A new study from the UK's Institute for Public Policy Research (IPPR) just exposed something most marketers haven't thought about yet: AI search engines are building a publishing hierarchy, and your brand's citation odds are being decided right now.

The headline finding: When ChatGPT responds to news queries, it cites The Guardian in 58% of answers. Reuters appears in 40%. The Independent in 35%.

Meanwhile, The Telegraph shows up in just 4% of responses. GB News in 3%. The Sun in 1%. And the BBC—the UK's most trusted news outlet—appears in 0% of ChatGPT answers.

Zero.

This isn't a bug. It's the new gatekeeping layer of AI search—and it's reshaping which publishers (and by extension, which brands mentioned in those publishers) get discovered by millions of users asking questions to ChatGPT, Perplexity, and Gemini.

If you're investing in earned media to drive awareness, this research should fundamentally change your publication strategy.

$2

The Institute for Public Policy Research tested four major AI tools—ChatGPT, Google Gemini, Perplexity, and Google AI Overviews—across 100 randomly generated news queries.

They analyzed over 2,500 citation links to see which publications AI engines favor.

The results were stark:

$3

- The Guardian: 58% of answers

- Reuters: 40%

- The Independent: 35%

- Financial Times: Frequently cited

- The Telegraph: 4%

- GB News: 3%

- The Sun: 1%

- Daily Mail: 0%

- BBC: 0%

$3

- BBC: 52.5% (completely absent from ChatGPT)

- Mixed use of Guardian, Reuters, and regional sources

$3

- BBC: 36%

- The Guardian: Frequently cited

- Higher diversity than ChatGPT

$3

- The Guardian: 53% (similar to ChatGPT)

- Limited BBC presence

Here's what's important: Each AI platform has a different publishing hierarchy. And those hierarchies aren't based purely on trust, quality, or readership—they're shaped by licensing deals, content access agreements, and editorial filtering decisions made by AI companies.

$2

The BBC is the UK's most popular and most trusted news source. It has massive reach, editorial credibility, and government backing.

So why does ChatGPT cite it 0% of the time?

Because the BBC blocked OpenAI from accessing its content.

Last year, the BBC threatened legal action against Perplexity for using BBC articles without permission. OpenAI respected the BBC's robots.txt directives and stopped crawling its content.

Result: The UK's most trusted journalism is invisible to the world's most popular AI tool.

Meanwhile, Google AI Overviews do cite the BBC (52.5% of news queries)—because Google has a licensing agreement with the BBC.

This is the new gatekeeping mechanism: AI companies decide which publishers to cite based on business relationships, not just editorial quality.

And for brands trying to earn citations, this creates a new strategic layer: You need to appear in the publications that AI engines actually index and prioritize.

$2

What the IPPR data reveals is a citation concentration problem—a handful of publishers dominate AI responses, while the long tail of media gets functionally ignored.

Compare the distribution:

- Top 3 publishers on ChatGPT: 58% + 40% + 35% = 133% (overlap, but dominance is clear)

- Bottom-tier publishers: 0-4%

This mirrors early SEO dynamics: Google's algorithm created a "rich get richer" effect where top-ranking sites accumulated backlinks, authority, and traffic—while page-two results got almost nothing.

AI citation works the same way. If ChatGPT habitually cites The Guardian for news queries, it will:

1. Reinforce The Guardian's authority in its training data

2. Drive more user attention to Guardian-cited sources

3. Create a feedback loop where The Guardian becomes the default news source for AI answers

For brands, this means:

Getting mentioned in The Guardian (or Reuters, or Financial Times) is exponentially more valuable than getting mentioned in a publication ChatGPT doesn't cite.

$2

Most brands approach earned media with a "more is better" mentality:

- Publish bylines on Medium, LinkedIn, and niche blogs

- Chase any press mention, regardless of publication tier

- Measure success by raw placement volume

The IPPR research shows why that strategy is failing in the AI era.

It's not about volume. It's about citation authority.

Here's the new earned media hierarchy:

$3

These publications appear in AI responses consistently across platforms:

- The Guardian (58% ChatGPT, 53% Gemini)

- Reuters (40% ChatGPT)

- Financial Times (frequent ChatGPT)

- BBC (52.5% Google AI, 36% Perplexity)

- Bloomberg, WSJ, NYT (likely high, though not in IPPR's UK-focused study)

Value: Mentions in these publications drive AI citations, which drive discovery, which drive brand awareness.

Barrier: High editorial standards, established relationships, or performance-based PR infrastructure.

$3

These publications get cited sometimes, but not reliably:

- The Telegraph (4% ChatGPT)

- Regional news outlets

- Industry trade publications (TechCrunch, VentureBeat, Forbes)

Value: They contribute to citation diversity and can drive direct traffic, but they're not AI-dominant.

Barrier: Moderate. Accessible via byline programs, PR pitching, or contributor networks.

$3

These publications don't appear in AI citations at all:

- Personal blogs

- Medium posts

- LinkedIn articles

- Low-authority niche sites

- Publications that block AI crawlers

Value: They may drive referral traffic or backlinks, but they don't contribute to AI visibility.

Barrier: Low. Accessible to anyone, which is why they don't carry AI authority.

If your earned media strategy focuses on Tier 3 placements, you're building visibility in a discovery layer that AI engines ignore.

$2

The IPPR study points to a critical factor most brands miss: AI companies are cutting licensing deals with select publishers, which determines citation access.

Examples:

- OpenAI has a licensing deal with Axel Springer (Politico, Business Insider)

- Google has partnerships with Reuters, AP, and the BBC

- Perplexity has ongoing negotiations (and legal battles) with major outlets

These deals create a paid moat around AI citations. If your brand appears in a publication that doesn't have a licensing deal with OpenAI or Google, your citation odds drop—even if the article is high-quality.

This is the equivalent of Google deciding that only certain websites can rank in the top 10 search results, regardless of SEO optimization.

For brands, this means:

You can't brute-force AI visibility. You need placements in publications that AI companies have indexed and prioritized.

$2

The IPPR study was motivated by public policy concerns: if AI tools only cite a narrow range of publishers, users are exposed to fewer perspectives, which harms democratic discourse.

IPPR's recommendations:

1. Require AI companies to pay publishers fairly - Collective licensing agreements

2. Introduce "nutrition labels" for AI news - Show users where citations come from

3. Fund public-interest AI news - Support BBC-led AI news services

These proposals are aimed at preserving media diversity. But they also reveal a market reality:

The current AI citation hierarchy is unstable. It's based on early-stage licensing deals, crawler access decisions, and editorial filtering—not a mature, standardized framework.

This is both a risk and an opportunity:

- Risk: If you're over-indexed on publications that AI engines deprioritize, your earned media ROI collapses.

- Opportunity: If you establish citation authority now in Tier 1 publications, you build defensible AI visibility before the market consolidates.

$2

Here's a tactical exercise to assess whether your current earned media strategy is AI-ready:

$3

Pull your earned media tracker and list the last 10 bylines, press mentions, or podcast appearances.

$3

For each publication, run this test:

1. Ask ChatGPT a news query related to your category (e.g., "What are the top earned media platforms for B2B?")

2. Check if ChatGPT cites that publication in its response

3. Repeat for Perplexity and Google Gemini

$3

- Tier 1 (AI-cited consistently): 3 points

- Tier 2 (AI-cited occasionally): 1 point

- Tier 3 (AI-never cited): 0 points

$3

- Total points / 30 = Citation Authority %

- 70%+: Your earned media strategy is AI-optimized

- 40-69%: Mixed. You're getting some citations, but not dominant

- <40%: You're investing in publications that AI engines ignore

If you're below 70%, you need to shift budget toward Tier 1 placements.

$2

Here's the problem with traditional PR:

You pay a retainer ($5k-$15k/month). Your PR firm pitches journalists at top-tier outlets. Sometimes they get placements. Sometimes they don't. You can't predict which publications will say yes.

If you're optimizing for AI citations, this model fails—because you need guaranteed access to Tier 1 publications, not "best effort" pitching.

The alternative is performance-based earned media:

1. You identify target publications (The Guardian, Reuters, Forbes, TechCrunch)

2. Your earned media partner guarantees placements in those outlets

3. You only pay for delivered placements, not hours or retainers

4. Each placement is AI-citation-optimized (expert quotes, data citations, category positioning)

This is how you de-risk earned media in the AI era. Instead of hoping your PR firm lands a Guardian byline, you contract for it.

$2

If the IPPR data alarms you (and it should), here's your action plan:

$3

- Run the AI citation test above

- Identify which publications AI engines cite

- Shift budget toward Tier 1 placements

$3

- The Guardian, Reuters, Financial Times (ChatGPT-friendly)

- BBC, AP (Google AI-friendly)

- WSJ, Bloomberg, NYT (likely high-citation across platforms)

$3

- After every Tier 1 placement, test 20 related queries in ChatGPT/Perplexity

- Document whether your brand's citation frequency increases

- Build a case study: "Forbes byline → 3x ChatGPT citations"

$3

- Use tools like RankShift.ai, Brandlight, or AnswerManiac

- Track your brand's citation frequency monthly

- Compare against competitors

$3

- If a publication never appears in AI responses, it's not building citation authority

- Redirect that budget to Tier 1 placements

$2

The IPPR study is a wake-up call: AI companies are deciding which publishers get cited, and those decisions are shaping which brands get discovered.

The Guardian's 58% citation rate isn't an accident. It's the result of:

- Licensing agreements with OpenAI

- High editorial quality that AI models trust

- Content accessibility (no paywalls for AI crawlers)

- Topical relevance to user queries

For brands, this creates a new strategic imperative: Earned media isn't just about getting press anymore. It's about getting press in publications that AI engines cite.

If your brand appears in 50 niche blogs but never in The Guardian, Reuters, or Financial Times, you're invisible to AI search—no matter how much content you publish.

The citation hierarchy is forming right now. The brands that establish authority in Tier 1 publications will own AI visibility for the next decade.

The ones that keep chasing vanity press mentions in Tier 3 outlets will wonder why their brand never appears in ChatGPT answers.

$2

The IPPR data reveals a bifurcating market:

- Winners: Brands mentioned in AI-citation-dominant publications

- Losers: Everyone else

There's no middle ground. ChatGPT doesn't cite "a little bit." It either includes you in its answer—or it doesn't.

The question for your earned media strategy is simple:

Are you investing in publications that AI engines cite?

If yes, you're building citation authority—the AI-era equivalent of domain authority.

If no, you're renting visibility in a discovery layer that's being phased out.

The gatekeepers have changed. The measurement has changed. The strategy has to change too.

---

Jaxon Parrott is CEO and founder of AuthorityTech, the earned media engine for brands that need guaranteed AI visibility. AuthorityTech delivers Tier 1 placements in publications like Forbes, TechCrunch, and VentureBeat—no retainers, just results. Book a strategy call to see how we guarantee citations in the publications that AI engines actually cite.