Will Google's Core Update Hurt Your AI Citations?

Every time Google confirms a core update, the same email lands in my inbox by day three. Rankings are jumpy, traffic dipped, and someone wants to know if the AI visibility work we just funded is about to wash out with it.

The May 2026 core update is the latest trigger. Google confirmed it started rolling out on May 21 and it finished around June 4, per Search Engine Land's tracking. Early reads called it a high-amplitude update, on par with the March 2026 one for sheer volatility.

So here is the honest answer to the question everyone is actually asking: a Google ranking drop does not automatically cut your AI citations. Most of the time the two move on separate tracks. Sometimes they fall together, and the reason they fall together tells you exactly what to fix.

A ranking drop and a citation drop are two different events.

Core update vs AI citations

A core update and an AI engine score different things

A Google core update re-judges which pages deserve their ranking. AI engines judge which passages are safe to reuse in an answer, retrieved from a far wider source pool. The two events overlap only where content quality is genuinely weak. Source: Search Engine Land coverage of the Google May 2026 core update.

A core update and an AI engine score different things

What you are looking at	Ranking layerGoogle core update	Retrieval layerAI citation layer
What you are looking atWhat it scores	Ranking layer · Google core updatePage-level relevance and trust against a specific query.	Retrieval layer · AI citation layerPassage-level usefulness and provenance, pulled from a much wider pool than Google's top 10.
What you are looking atWhat moves on a rough update	Ranking layer · Google core updateYour SERP position, your organic click volume, your impressions.	Retrieval layer · AI citation layerOften nothing mechanical. Citations get re-evaluated on their own clock, not Google's.
What you are looking atThe shared failure mode	Ranking layer · Google core updateThin, mass-produced pages with no human editing dropped hardest in May 2026.	Retrieval layer · AI citation layerThose same thin pages were already weak grounding sources. Here both channels fall together.
What you are looking atThe signal Google ignores	Ranking layer · Google core updateOff-site mentions, reviews, and third-party comparisons do not set your ranking.	Retrieval layer · AI citation layerThose off-site signals carry real weight in which brands a model decides to cite.
What you are looking atThe self-inflicted risk	Ranking layer · Google core updatePanic-rewriting pages mid-update can deepen a ranking drop.	Retrieval layer · AI citation layerEditing a page AI already cites can delete the exact passage that earned the citation.

This post walks through what the May 2026 update changed, why your AI citations usually survive a ranking shake-up, the cases where they genuinely do not, and what to do while the dust settles.

What the May 2026 core update actually changed

A core update is Google re-judging which pages deserve their ranking. It does not target AI citations at all. The May 2026 version hit one type of content hardest, and it surfaced a data point that should change how you read your own dashboards.

The update ran from May 21 to roughly June 4

Google confirmed the rollout on May 21 and it completed in early June. That two-week window is normal for a core update, and the volatility inside it is also normal. Rankings swing, partially recover, and swing again before settling.

Google's own guidance is to wait. It tells site owners not to draw conclusions until a week after the update finishes, which puts a clean read somewhere in the week of June 8. The Search Central documentation on core updates has said this for years, and it still holds.

AI-generated thin content took the hardest hit

The clearest pattern in the first week was which sites dropped. Thin informational pages, mass-produced content with no real human editing, affiliate and review aggregators, and shaky YMYL pages absorbed the steepest losses, per the day-five volatility read from Digital Applied.

That detail matters more than it looks. The content Google punished here is the same content AI engines were already reluctant to cite. Cheap, generic pages are weak in both systems for the same underlying reason: they do not say anything an engine can trust enough to reuse.

Brands cited in AI Overviews saw clicks rise, not fall

Here is the anchor most coverage buried. During this update, brands cited in AI Overviews reported higher volumes of both organic and paid clicks than brands that were not cited. Being in the AI answer correlated with more traffic on both channels, not less.

That flips the usual core-update story. Most update coverage is defensive: do not panic, do not react. This says the brands already winning AI citations had a buffer the uncited ones did not. We covered the click side of this in why cited brands get 2.3x more AI Overview clicks.

Being in the AI answer was a buffer, not a bystander.

Why a Google ranking drop does not automatically cut your AI citations

This is the part that calms most clients down. Your Google ranking and your AI citation rate are produced by two different machines asking two different questions. A core update tunes one machine. It does not reach into the other.

The clearest way to see it is to put the two questions side by side.

A Google core update asks:

• •Does this page still deserve its position for the query?
• •Is the site trustworthy and useful at the page level?
• •Was this written for people, or churned out at scale?

An AI engine asks:

• •Can I extract a clean passage that answers the prompt?
• •Does an independent source back this claim?
• •Is this the freshest, clearest version of the answer?

Those lists barely overlap. That is why a page can slide three positions in Google and still get quoted verbatim by ChatGPT the same week. Five specific reasons drive the gap.

Reason #1: AI retrieves from a far wider pool than the top 10

Google ranking is a contest for ten visible slots. AI retrieval pulls from a much larger candidate set, often dozens of pages per sub-query, before it picks what to cite. Slipping from position 4 to position 8 barely changes whether you are in that retrieval pool.

We dug into this gap in why Google rankings no longer predict AI citations. The overlap between top-10 rankings and AI citations has been collapsing for a year. A core update nudges the rankings without touching most of that wider pool.

Reason #2: AI cites passages, your ranking measures whole pages

Google scores the page as a unit. AI engines lift a 40-to-60-word passage out of the page and judge that fragment on its own. A core update can downgrade a page's overall standing while the one clean answer block inside it stays just as quotable.

AI does not read your Google ranking. It reads your paragraph.

This is why structurally strong pages are resilient. If your answer is stated cleanly in two sentences near the top, that passage survives a ranking wobble because the model never needed the ranking to find it.

Reason #3: Off-site signals carry weight a ranking never captures

Google's core updates re-weigh on-page and site-level signals. They do not measure how often Reddit, G2, a review site, or an industry roundup mentions you. AI engines lean heavily on exactly those third-party signals when they decide which brand to name.

So a chunk of your AI visibility lives in places a core update cannot reach. Off-site authority is the part of your presence that does not show up in Search Console and does not move when Google re-tunes its ranking model.

Reason #4: Five engines retrieve differently, so one SERP drop is not five citation drops

You have one Google ranking. You have separate citation behavior in ChatGPT, Claude, Gemini, Perplexity, and AI Overviews. A Google core update directly touches one of those five surfaces, and only the grounding layer of that one.

A ranking drop is a single-channel event. Your Claude and Perplexity citations were never indexed to your Google position, so they do not move because Google moved. We have watched this firsthand in cases where clicks dropped while AI absorption held steady.

Reason #5: Citation drift runs on its own clock, not Google's

AI citations churn constantly, with or without a Google update. The average AI citation has a half-life measured in weeks, and sources rotate in and out on a schedule that has nothing to do with Google's release calendar.

That means if your citations shift the same week as a core update, the update may not be the cause at all. We mapped this background churn in why your AI visibility changes weekly. Correlation in timing is not proof of cause.

When a core update does threaten your AI citations

The reassurance has limits. There are real cases where a core update and an AI citation loss are the same event wearing two faces. If any of these describe your pages, the two channels are linked and you should treat the ranking drop as a warning about your AI visibility too.

If your pages were thin AI content, both channels fall together

The May 2026 update punished cheap, generic, machine-written pages. Those pages were already poor grounding sources. If a page lost its ranking because it reads as filler, it was probably losing AI citations on a slower timeline anyway.

This is the one case where the core update is a useful diagnostic. A ranking collapse on a thin page is Google confirming what the models had already decided. We laid out the mechanism in whether AI content is hurting your AI search visibility.

If AI grounded its answer on your ranked page, the answer can move

AI Overviews and ChatGPT search both ground some answers in live results that lean on Google's index. When your page falls out of the set the engine retrieves for a query, the grounded answer can quietly swap in a competitor.

This is a narrower risk than it sounds, because it only affects the live-grounded share of answers, not the citations drawn from training data or off-site sources. But for high-intent, freshness-sensitive queries, a ranking drop can ripple into the AI answer.

If you panic-edit cited pages, you cause the drop yourself

This is the most common self-inflicted wound. Rankings dip, a team rushes to rewrite the affected pages, and in the rewrite they delete or bury the exact passage AI had been citing. The core update did not remove the citation. The reaction did.

Panic is the only part of a core update you fully control.

A page AI already cites is an asset. Rewriting it mid-update without checking what was getting quoted is how brands turn a recoverable ranking wobble into a permanent citation loss.

What to do while the update settles

The right response to a core update is mostly disciplined waiting plus a few targeted checks. Here is the sequence I give clients in the week an update completes.

Step 1: Freeze reactive content edits until the update finishes

Do not rewrite anything during the rollout. Rankings are still moving, so any change you make now is impossible to attribute. With the May 2026 update completing around June 4, hold edits and let the volatility settle before you touch a page.

Step 2: Separate your Google data from your AI data before you diagnose

Pull Google Search Console and your AI citation tracking into two separate views. If rankings dropped but citation share held, you have a ranking problem, not an AI problem, and the fixes are different. Most teams blur the two and end up treating a SERP dip as an AI emergency. The AI search measurement guide covers how to set up the split.

Step 3: Re-baseline AI citations the week of June 8

Google says wait a week after completion before judging anything. Apply the same patience to AI. Re-run your prompt set in the week of June 8 and compare citation share against your pre-update baseline. One clean read beats five anxious daily checks during the swing.

Step 4: Audit whether your cited pages read as thin or machine-written

For any page that lost rankings, ask the harder question: would a model trust this enough to quote it? If the page is generic filler, the core update and your weak AI citations share one root cause, and the fix is rebuilding the page with real substance, specific claims, and clean answer blocks.

Step 5: Strengthen the off-site signals a Google update cannot touch

The most update-proof part of your AI visibility lives off your own domain. Reviews, third-party comparisons, community threads, and earned mentions feed AI citations and sit entirely outside Google's ranking model. Investing there builds visibility that no core update can take back.

FAQ

Does a Google core update directly affect AI citations?

No. A core update re-judges page rankings in Google's own results. It does not target ChatGPT, Claude, Gemini, or Perplexity citations, which are produced by separate retrieval systems. It can affect the live-grounded share of AI Overviews answers, but most AI citations are drawn from a wider source pool and off-site signals that a core update never touches.

My rankings dropped in the May 2026 update. Will my AI visibility drop too?

Usually not by the same mechanism. If your page is solid and just lost a few positions, your AI citations are likely intact because AI retrieves from a far wider pool than the top 10. The exception is thin, machine-written pages, which were hit hardest by this update and were already weak AI sources. In that case both channels reflect the same quality problem.

Should I rewrite pages that lost rankings during the update?

Not during the rollout, and not before you check what AI was citing on those pages. Rewriting a page mid-update makes the result impossible to attribute, and deleting a passage AI already quotes can cause a citation loss the update never would have. Wait until completion, separate ranking data from citation data, then edit only what the data tells you to.

How long should I wait before judging the damage?

About a week after the update completes. The May 2026 update finished around June 4, so a clean read lands in the week of June 8. Re-run your AI prompt set then and compare citation share against your earlier baseline, rather than reacting to daily swings during the volatile rollout window.

Why did some brands gain clicks during the core update?

Brands cited in AI Overviews reported higher volumes of both organic and paid clicks than uncited brands during this update. Being in the AI answer acted as a buffer. It is the clearest sign that earning AI citations is a click-positive investment, not just a defensive one.

The bottom line

The fear behind the question is that AI visibility is just SEO with a new label, so a Google update would sink both at once. The May 2026 data says otherwise.

A core update re-judges your Google rankings. Your AI citations are scored by different systems, retrieved from a wider pool, weighted toward passages and off-site signals, and churning on their own clock. The two channels only fall together when the underlying content is genuinely thin, which is a content problem the update merely exposed.

So separate the two before you react. Confirm whether you have a ranking drop, a citation drop, or just normal volatility. Then fix the one you actually have. The brands that came through this update strongest were not the ones with the best Google positions. They were the ones already cited in the AI answer, where a core update could not follow.