Should Law Firms Disclose When AI Has Been Used to Create Content?

AI use in law firm content is no longer unusual

A lawyer writes an article and uses AI to improve the structure. Similarly, their marketing team uses it to suggest alternative headlines.

A writer uses it to shorten a paragraph, and a law firm asks an AI tool to produce the first draft of an entire client alert.

A whole other firm uses automation to generate dozens of legal updates from publicly available information.

All of these may be described as “using AI” but they are clearly not the same thing.

This observation makes a difference and matters when asking whether law firms should tell readers that AI has been involved in producing their content.

The question is no longer: “Was AI used?”

The better question today is: “What role did AI actually play in creating what the reader is seeing?”

Using AI does not automatically mean a disclosure is required

There is an important difference between AI-assisted content and content that has been substantially generated by AI.

Consider the difference:

Treating every one of these situations in exactly the same way makes little sense.

A lawyer who uses AI to correct punctuation has not suddenly stopped being responsible for the article but if an entire article is generated through automation and published under the name of a lawyer who has contributed very little to it, the transparency question becomes much more important.

Google does not treat all AI content as automatically problematic

There is sometimes an assumption that AI-generated content must be disclosed because otherwise Google will penalise the website.

That is not how Google's published guidance works.

Google's focus remains primarily on whether content is useful, original, trustworthy and created for people rather than simply to manipulate search rankings. It also says that AI or automation disclosures can be useful where a reader might reasonably wonder how the content was created, particularly where automation has played a substantial role.

This has seemed to become relevant for law firms while creating content.

The SEO doesn't ask anymore: “Did we tell Google that AI helped us?”

The more important questions are:

• Is the article genuinely useful?

• Is the information accurate?

• Is there real legal expertise behind it?

• Does it add anything beyond information already available elsewhere?

• Is the named author genuinely responsible for what has been published?

• Was AI used to assist expertise or to replace it?

AI in itself is not the shortcut to good content but an AI disclosure does not turn poor content into good content.

The author name creates an expectation

This is particularly important for law firms because legal articles often carry the name of an individual lawyer.

Imagine an article that says: By Sarah Williams, Partner – Competition Law

A reader will naturally assume that Sarah's professional knowledge, judgment or perspective sits behind what they are reading. That does not necessarily mean Sarah personally typed every sentence.

Law firms have always used editors, researchers, marketing teams, external writers and communications professionals to help produce content.

The important issue is whether the final article genuinely represents the lawyer's knowledge and has been reviewed and approved by the person whose expertise is being presented.

AI introduces the same question in a new form.

If it helped Sarah organise her ideas, improve readability or turn her notes into a polished article, that is very different from automatically generating an article about competition law and placing Sarah's name above it without meaningful involvement from her.

The byline should still mean something.

The real issue is responsibility

The most important question may therefore be simpler than disclosure like "Who is responsible for the final content?"

If a law firm publishes an article explaining a legal development, someone within the firm should be prepared to stand behind:

• the legal position stated;

• the cases and authorities referred to;

• the interpretation being offered;

• the practical implications suggested; and

• the overall accuracy of the article.

AI cannot take professional responsibility for an incorrect statement, nor can it be held accountable because a case citation was invented.

It also, cannot explain to a client why incorrect information appeared under a partner's name.

Professional guidance on lawyers' use of generative AI has increasingly emphasised human oversight, accuracy, confidentiality and professional responsibility rather than treating AI as an independent decision-maker.

For examples, in the US, the ABA's Formal Opinion 512, stresses existing duties such as competence, confidentiality, communication and supervision, while the SRA's August 2026 warning similarly emphasises appropriate human review and the risk of AI-generated inaccuracies.

The tool may assist but responsibility always stays with the lawyer and the firm.

Disclosure becomes more important as AI's contribution becomes larger

A useful way of approaching disclosure is to think of it as a scale rather than a yes-or-no rule.

It is essential to know that AI assistance and AI authorship are not the same thing.

Human review should mean more than proofreading

A law firm should also be careful with the phrase:

“Reviewed by a lawyer.”

Reviewing AI-generated legal content should not simply mean correcting grammar and making the article sound professional.

A meaningful review may require checking:

• whether the cited judgment actually exists;

• whether the court said what the article claims;

• whether legislation is current;

• whether quotations are accurate;

• whether the relevant jurisdiction has been identified correctly;

• whether an exception has been overlooked;

• whether the conclusion goes further than the authority supports; and

• whether the practical advice reflects how the law actually operates.

Generative AI can produce information that sounds convincing even when it is wrong. As a result, a polished sentence is not evidence of an accurate legal proposition.

The confidentiality question is different from the publishing question

There is another AI issue that should not be confused with disclosure to readers, and that is a lawyer may use AI while preparing public content without entering any confidential information.

But they may also paste:

• client emails;

• internal legal advice;

• facts from an ongoing matter;

• unpublished agreements;

• confidential documents; or

• identifiable client information

into an external AI system while preparing that content.

At that point, the question becomes much larger than whether the finished article should carry a note saying “AI assisted”.

The firm needs to consider how the tool handles data, what information may be retained, what its professional obligations require and whether the information should have been entered into the system at all.

Disclosure at the bottom of an article does not solve a confidentiality problem that occurred during its creation.

What about ghostwriting?

AI also raises an interesting comparison with something law firms have done for years.

• Partners do not always write every word appearing under their names.

• A marketing professional may interview a lawyer.

• An external writer may prepare the first draft.

• A junior lawyer may conduct research.

• An editor may substantially restructure the article.

Yet firms do not normally publish: “This article was drafted by our marketing team and edited by the partner.”

Why should AI automatically be different? Sometimes it should not be.

The more useful principle is whether the content honestly represents the knowledge and views attributed to the named lawyer or firm.

1. A human writer can produce an article that a partner barely reviews.

2. AI system can help turn an hour-long conversation with a partner into a piece that accurately represents the partner's thinking.

The first may involve no AI whatsoever and still create a questionable impression of authorship. The second may involve substantial AI assistance while remaining genuinely grounded in the lawyer's expertise.

The disclosure itself should say something useful

If a firm decides disclosure is appropriate, a vague label such as: “Created using AI” may create more questions than it answers.

• Did AI correct the grammar?

• Did it write one paragraph?

• Did it conduct the research?

• Did it produce the entire article?

• Was the content reviewed by a lawyer?

A more meaningful disclosure could explain the role AI actually played.

For example:

“AI tools were used to assist with editing and structure. The legal analysis was prepared and reviewed by the author.”

Or, where the involvement was greater:

“Generative AI was used to assist in preparing an initial draft. The legal authorities, analysis and final text were reviewed and approved by the firm's legal team.”

The point of disclosure should be to give the reader useful information rather than simply place an AI label somewhere on the page.

Law firms may need an AI content policy before they need an AI label

The bigger issue for many firms may not be what sentence appears underneath an article but what happens before publication.

A sensible internal content process could answer questions such as:

• Which AI tools may lawyers and marketing teams use?

• What information may never be entered into them?

• Can AI conduct legal research for published content?

• Who verifies cases and authorities?

• How much human review is required?

• Can AI-generated material appear under a lawyer's byline?

• When should AI use be disclosed?

• Who gives final publication approval?

• How should the firm deal with corrections?

If those questions have not been answered internally, adding a disclosure statement to the website addresses only a very small part of the issue.

Readers are ultimately relying on the firm, not the tool

A client reading a law firm's article is not choosing the firm's AI system but whether to trust the firm.

That trust comes from believing that the content represents genuine legal knowledge, has been checked carefully and has been published by people willing to take responsibility for it.

AI can help lawyers produce content faster like:

• It can help organise ideas.

• It can improve structure.

• It can make existing writing easier to read.

• It can help turn a partner's scattered observations into something publishable.

None of that automatically makes the content less valuable. The problem begins when technology creates the appearance of expertise that the firm itself has not actually contributed to.

So, should law firms disclose AI use?

Sometimes, but “Was AI used?” is probably too crude a test.

A better set of questions is:

• How substantially did AI contribute?

• Did it influence the legal substance or merely the presentation?

• Has a qualified person properly reviewed the output?

• Is the named author genuinely responsible for the article?

• Would the absence of disclosure give the reader a misleading impression?

• Would a reasonable reader expect to know how this content was produced?

If AI merely helped a lawyer communicate genuine expertise more clearly, a disclosure may add very little, and if AI largely created the substance that is being presented as the work or thinking of a lawyer, transparency becomes much harder to dismiss.