Your law firm client’s traffic dropped after a core update. You’ve been running AI-generated blog posts for six months. The instinct is to connect those two facts. Before you restructure a content workflow based on that instinct, check whether the data actually supports it.
What Did the March 2024 Core Update Actually Do to AI Content Rankings?
The Helpful Content Update, as a standalone ranking classifier, no longer exists. Google incorporated the helpful content system into its core ranking algorithm in March 2024. Agencies still strategizing around “HCU recovery” are targeting a mechanism that was deprecated over two years ago.
What this means in practice: sites that lost visibility under the old HCU signals do not recover through a separate recovery pathway. They recover, if they do, through standard core update cycles. The question is not “when does my client get out from under HCU?” The question is whether the content meets the current core quality standard.
That standard has not changed materially. Danny Sullivan addressed it directly on the December 2025 Search Off the Record podcast, with his comments on what Google’s ranking systems reward confirming that the evaluation standard for AI search is identical to the standard for traditional search: is the content designed for people?
What Does Google2019s Current Spam Policy Say, and Where Is the Line for Agencies?
The February 2023 Google blog post on AI content guidelines is the most-cited authority in every AI content debate. It predates the May 2024 spam policy revision by 15 months. Agencies quoting “it’s just content” from that post are citing a statement that has since been operationalized with significantly more specific policy language.
The violation is defined in Google’s scaled content abuse spam policy as generating “many pages for the primary purpose of manipulating search rankings and not helping users.” The specific example given: “Using generative AI tools or other similar tools to generate many pages without adding value for users.”
That’s not a penalty for using AI. That’s a penalty for volume without value. The distinction matters because fixing the wrong thing produces no result.
The operative words are “many pages” and “without adding value.” Volume paired with genuine legal depth, jurisdiction-specific detail, and attorney review is not the pattern the spam policy targets. Volume paired with thin, templated output is. For agencies operating in competitive legal markets, attorney SEO in the AI era operates under enforcement mechanisms the 2023 blog post alone does not prepare you for.
What Does the Ranking Data Actually Show About AI Content at Position 1?
Two large-scale studies produce findings that look like they contradict each other until you read them carefully.
A Semrush analysis of 42,000 blog posts, published in their April 2026 ranking study, found that human-written content holds Position 1 eighty percent of the time. Pure AI content holds Position 1 nine percent of the time. In the same study, 72% of the 224 SEO professionals surveyed said AI content ranks as well as human content. Perception and data are in direct conflict.
Ahrefs published findings from a 600,000-page ranking analysis in July 2025: the correlation between AI content percentage and ranking position is 0.011, effectively zero. 86.5% of top-ranking pages contain some AI content. Pure AI rarely reaches the top position, but AI-assisted content is nearly ubiquitous across Page 1.
Both studies used AI detection tools to classify content. Both vendors sell AI detection products. The data is real. The framing deserves scrutiny.
The reconciled read: AI content does not categorically hurt rankings. Pure AI content rarely wins the top position for competitive queries. The difference between ranking and winning is where most agency economics break down.
For law firm content, the conditions compound. YMYL treatment means quality signals carry more weight in competitive practice areas than they do in non-legal verticals. A personal injury blog post does not operate in the same ranking environment as a lifestyle or general business content piece.
Where Does the Human Edit Layer Change SEO Outcomes for Law Firm Content?
The Ahrefs and Semrush data converge on one implication: AI-assisted content performs comparably across Page 1 broadly but underperforms at Position 1 for competitive queries. Four specific interventions define what the edit layer actually requires in legal content. None of them are generic copyediting functions.
Legal accuracy verification is the first. Detection scores measure statistical probability of model generation, not factual accuracy. Agencies that have reviewed the limits of AI detection in legal content understand why factual review by someone who knows the law is not optional in this vertical. A jurisdiction-specific error does not show up in a detection score.
Jurisdiction-specific depth is the second. “California’s two-year statute of limitations for personal injury claims” is more useful than “deadlines vary by state.” The useful version is also the version that appears in AI Overviews. Generic legal summaries are the content type AI produces most efficiently and the content type that positions least competitively.
E-E-A-T attribution is the third. Google’s current documentation on people-first content standards explicitly asks whether AI use is self-evident to visitors through disclosures. An attorney byline with a bar number sends a different signal than a generic “legal team” attribution, for both Google and for the prospective client reading the page.
Topical originality is the fourth. A piece that takes a position, cites a named source, and moves the reader’s understanding forward outperforms neutral summaries of existing law for competitive queries. Neutral legal summaries are the output AI produces most confidently and positions least effectively.
How Do You Measure Whether AI Content Is Hurting a Specific Client’s Rankings?
Before attributing a ranking drop to AI content, map the timing against Google’s core update calendar. The March 2024 core update, the June 2025 core update, the March 2026 core update, and the May 2026 core update currently rolling out are the relevant correlation points. A traffic drop that aligns with a specific update date narrows the diagnostic considerably.
Second, distinguish between Position 1 share and Page 1 presence. The Semrush data shows AI content clustering in positions 2 through 4, not absent from Page 1. If a client maintains Page 1 presence but loses Position 1 to competitors with more substantively edited content, the problem is the edit layer, not AI use.
Third, isolate AI Overview citation share from traditional organic rankings. AI Overview clicks, as documented in Google’s guidance on AI Overviews and site owners, appear in Search Console under the “Web” search type, making direct attribution to AI Overview presence difficult. Treat AI Overview visibility as a separate performance signal with its own content quality logic.
Fourth, compare engagement metrics on AI-produced versus attorney-reviewed pages covering equivalent topics. Session duration and bounce rate differences between content types give a practical quality signal independent of detection scores.
Common Questions About AI Content and Law Firm SEO
Does Google penalize AI-generated content?
Google does not penalize AI content as a category. Its scaled content abuse policy penalizes generating many pages without adding value for users, with AI tools cited as one mechanism for that violation. The trigger is volume without value, not AI use itself.
What is the current status of the Helpful Content Update?
The Helpful Content Update as a standalone classifier was incorporated into core ranking in March 2024. There is no separate HCU recovery pathway. Sites improve through core update cycles by producing substantively better content.
How does legal content differ from other niches for AI SEO risk?
Legal content receives YMYL treatment, which increases the weight of quality signals in competitive practice areas. Named attorney authorship, jurisdiction-specific accuracy, and E-E-A-T attribution matter more for personal injury or criminal defense content than for non-YMYL verticals. The agency risk profile is higher.
Should agencies disclose AI use to law firm clients?
Google’s people-first content framework recommends disclosure where users might ask how content was created. For law firm clients, a separate framework governing what attorneys owe clients when AI writes content applies at the bar level, creating a stricter disclosure obligation than Google’s guidance requires.
Agencies That Have Been Inside This Problem Since Before AI Existed
We have produced attorney-reviewed content at scale for more than 300 law firms across North America. Lexicon is attorney-owned since 2012, with both founders holding JDs from Saint Louis University School of Law, and attorney review has been part of every workflow we’ve built since we opened. That’s not an AI-era compliance addition. It’s how the agency was designed. If your agency needs a content partner whose production sample spans more than a decade of Google updates, call us at 877-486-8123 or contact us online.
David Arato, JD, is the founder of Lexicon Legal Content, an attorney-owned legal content marketing agency serving law firms since 2012. He has spent over a decade managing AI-assisted content workflows across hundreds of law firm websites, tracking how Google’s quality updates intersect with legal content ranking performance. David is a frequent contributor to Attorney at Law Magazine and Attorney at Work and a recurring guest on legal marketing podcasts.
