Platform Liability and Section 230 Toolkit

By Casey Scott McKay · 2025-10-09

Section 230 is narrower than its critics assume and broader than most complaints allow for, and the difference between those two propositions decides whether a platform case ends on a motion to dismiss. This toolkit works the three elements: an interactive computer service, treatment as a publisher or speaker, and information provided by another. It sets out the material contribution test from Roommates.com that determines when a platform becomes a content provider itself. It covers the subsection (c)(2) moderation immunity and the good faith limit that constrains it. It then works the statutory exceptions for federal criminal law, intellectual property, and FOSTA. It closes with the claim theories that sidestep the statute entirely through product design, failure to warn, and contract.

IP and Technology > Internet | Toolkit | Published 9 October 2025 - Updated 28 October 2025 | Casey Scott McKay - marksy.us

Summary. Section 230 is narrower than its critics assume and broader than most complaints allow for, and the difference between those two propositions decides whether a platform case ends on a motion to dismiss. This toolkit works the three elements: an interactive computer service, treatment as a publisher or speaker, and information provided by another. It sets out the material contribution test from Roommates.com that determines when a platform becomes a content provider itself. It covers the subsection (c)(2) moderation immunity and the good faith limit that constrains it. It then works the statutory exceptions for federal criminal law, intellectual property, and FOSTA. It closes with the claim theories that sidestep the statute entirely through product design, failure to warn, and contract.

Keywords: section 230 · interactive computer service · information content provider · publisher liability · Zeran · Roommates.com material contribution · good samaritan moderation · subsection c2 · Enigma v Malwarebytes · statutory exceptions · federal criminal law · intellectual property carve-out · Perfect 10 v CCBill · FOSTA · product design claims · Lemmon v Snap · failure to warn · Gonzalez and Taamneh · moderation records · early dismissal

Start Here

A marketplace receives a complaint. A seller listed a counterfeit product, a buyer was defrauded, and the buyer is suing the marketplace for negligence, misrepresentation, and unfair trade practices.

The marketplace's counsel reaches for Section 230, and the answer is: partly. The claims premised on the seller's listing content are barred. The claims premised on the marketplace's own representations about its verification programme are not. The trademark claim is outside the statute entirely. And the negligence claim, if pleaded as a defect in the marketplace's own design rather than as a failure to remove the listing, may survive.

That is what platform liability analysis actually looks like. Not a shield or its absence, but a claim-by-claim assessment of what each count is really premised on.

This toolkit answers three questions.

What does the statute cover? Three elements, each contested, with the third doing most of the work.
What falls outside it? Statutory exceptions, and claim theories that avoid the elements.
How is the defence positioned? Early, on the pleadings, supported by records that show what the platform did and did not do.

If you read only one thing, read The Twenty-Six Words and Their Limits. It works the elements and the exceptions in the order a motion has to address them.

The Three Elements

The provision. 47 U.S.C. § 230(c)(1) provides that no provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.

Element one — an interactive computer service. Defined broadly at 47 U.S.C. § 230(f) to include any information service, system, or access software provider that provides or enables computer access by multiple users to a computer server. Websites, apps, marketplaces, forums, hosting providers, and email services all qualify, and this element is rarely contested.

Element two — treated as the publisher or speaker. The claim must seek to hold the defendant liable for a publishing function: deciding whether to publish, withdraw, postpone, or alter content. Zeran v. America Online established the broad reading, holding that notice of defamatory content does not create liability for failing to remove it.

Element three — information provided by another. The content must originate with someone else. Where the platform is itself the information content provider — defined at 47 U.S.C. § 230(f) as a person responsible in whole or in part for the creation or development of the information — the statute does not apply to that content.

The third element is where the fights are. A platform that creates content, that requires users to supply unlawful content, or that materially contributes to its unlawfulness is a content provider as to that material.

The threshold question for any motion. For each count, identify the duty the plaintiff says was breached. If the duty is to publish or not publish third-party content, the statute applies. If the duty arises from the platform's own conduct, it may not.

Material Contribution

Fair Housing Council of San Fernando Valley v. Roommates.com supplies the test that has governed since.

What the platform did. Required users to answer questions about sex, sexual orientation, and family status through drop-down menus, then used the answers to filter listings.

The holding. By requiring the answers and structuring them, the platform was responsible in part for developing the allegedly unlawful content, and was a content provider as to it. The free-text "additional comments" field, by contrast, was protected — the platform did not require or shape what users wrote there.

The test. A platform is a content provider where it materially contributes to the alleged unlawfulness of the content, as distinct from providing neutral tools that users employ for their own purposes.

Neutral tools remain protected. Search functions, categorisation, filtering, and recommendation mechanisms that operate without regard to the unlawful character of the content are ordinary publishing functions.

Which means the design question is: does the tool require, encourage, or shape the unlawful element? A drop-down forcing a discriminatory preference contributes; a text box that a user fills with a discriminatory statement does not.

Editing is generally protected. Selecting, editing, and arranging third-party content are traditional publisher functions, and minor alterations do not convert a platform into a content provider.

Creating content is not. A platform's own posts, its own reviews, its own descriptions, and its own representations about its service are its own content, and Section 230 does not reach them.

Moderation Immunity

The provision. 47 U.S.C. § 230(c)(2) protects a provider from liability for voluntarily taking action in good faith to restrict access to material the provider considers obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, and for enabling others to restrict access to such material.

Why it matters separately. Subsection (c)(1) protects against liability for content left up; subsection (c)(2) protects against liability for taking content down.

In practice, (c)(1) does most of the work. Courts have frequently held that removal decisions are also publishing decisions protected by (c)(1), which avoids the good faith inquiry that (c)(2) requires.

Good faith is a real limit. Enigma Software Group USA v. Malwarebytes held that (c)(2)(B) does not immunise blocking decisions made for anticompetitive reasons, reading "otherwise objectionable" as constrained by the statutory context.

Which means moderation motivated by competitive advantage rather than content concerns is exposed.

Documentation matters. A moderation decision recorded against a published policy, with the reason and the reviewer, is defensible. One with no record invites the argument that the stated reason is pretextual.

First Amendment considerations run alongside. Moody v. NetChoice addressed state laws regulating platform content moderation and confirmed that editorial curation implicates expressive interests, though it resolved the cases on the facial-challenge posture rather than settling the substantive questions.

Practical guidance. Publish the policy, apply it consistently, record the decisions, and give notice where the terms promise it — because contractual promises about moderation are enforceable independently of Section 230.

The Statutory Exceptions

47 U.S.C. § 230(e) removes four categories from the immunity.

Federal criminal law. The statute does not impair enforcement of federal criminal statutes. It does not, however, create a private right of action, and courts have differed on whether a civil claim predicated on a federal criminal statute falls within the exception.

Intellectual property. The immunity does not limit any law pertaining to intellectual property. Perfect 10 v. CCBill construed this as limited to federal intellectual property law, holding that state right of publicity claims remain barred — a reading not universally adopted.

Which means trademark and copyright claims are outside the statute. For copyright, 17 U.S.C. § 512 supplies its own safe harbour with its own conditions. For trademark, there is no equivalent safe harbour, and platform liability turns on contributory infringement doctrine.

The Electronic Communications Privacy Act and similar state laws are preserved.

Sex trafficking. FOSTA added 47 U.S.C. § 230(e)(5), removing immunity for federal civil claims under 18 U.S.C. § 1591, for state criminal prosecutions where the conduct would violate that section, and for state prosecutions under 18 U.S.C. § 2421A.

The FOSTA carve-out is narrower than commonly assumed — it requires conduct meeting the federal standards, not merely an allegation involving trafficking — but its practical effect on platform moderation has been substantial.

Practical consequence. For any claim, check the exceptions before the elements. A trademark claim against a marketplace is not a Section 230 case at all; it is a contributory infringement case.

The state law point matters commercially. Section 230 preempts inconsistent state law, which is what makes the immunity uniform, and the exceptions are the defined places where state law operates.

Claims That Avoid the Statute

The productive plaintiff strategy is not to defeat the immunity but to plead around it.

Contract and promissory estoppel. Barnes v. Yahoo! holds that a specific promise to remove content, relied upon, supports a promissory estoppel claim that does not treat the platform as a publisher — it treats it as a promisor. Terms of service that make removal commitments create this exposure.

Product design. Lemmon v. Snap holds that a negligent design claim directed at a feature — there, a speed filter — is not a publishing claim, because the duty alleged concerns the platform's own product rather than third-party content.

Which has become the principal route. Design claims about recommendation systems, engagement mechanics, age verification, and safety features are pleaded as product defects rather than as content complaints.

Failure to warn. Doe v. Internet Brands permits a failure-to-warn claim based on knowledge the platform acquired outside its publishing function, because the duty does not arise from publishing.

The platform's own statements. Representations about verification, safety, screening, or authenticity are the platform's own content. A marketplace that advertises a verification programme is liable for misrepresentation about that programme.

Regulatory obligations on the platform's own conduct. HomeAway.com v. City of Santa Monica upheld an ordinance imposing duties on the platform's own booking conduct rather than on the listings.

Aiding and abetting. Twitter v. Taamneh resolved the substantive standard for aiding and abetting liability under the anti-terrorism statute without reaching Section 230, and Gonzalez v. Google was disposed of on that basis — leaving the algorithmic recommendation question largely unresolved.

Force v. Facebook had held that algorithmic arrangement and recommendation of third-party content is a publishing function within the immunity, which remains the prevailing view in the circuits that have addressed it.

Which is the live frontier. Whether recommendation systems are neutral tools or material contributions is where platform litigation is now concentrated, and the answer is unsettled.

Asserting the Defence

Move early. Section 230 is an immunity from suit as well as from liability, and its value depends on resolving the case before discovery.

A motion to dismiss is the vehicle. The elements are typically apparent from the complaint, and courts routinely resolve the defence on the pleadings where the face of the complaint establishes them.

Take the counts one at a time. For each, identify the duty alleged. Publishing duties are barred; independent duties are not. A single order can dismiss some counts and not others, and that is a good outcome.

Address the third element carefully. The plaintiff will allege material contribution, and the answer requires describing what the platform's tools actually do — which is a factual account the motion has to supply without converting to summary judgment.

Watch the conversion risk. Extensive factual material attached to the motion invites conversion. Where the elements cannot be established from the complaint and judicially noticeable material, an early summary judgment motion may be better.

Handle the exceptions expressly. If an intellectual property claim is pleaded, address it under its own framework — 17 U.S.C. § 512 for copyright, contributory infringement doctrine for trademark — rather than under Section 230.

Preserve the (c)(2) argument as an alternative for removal claims, with the good faith record.

Anticipate the design-defect reframing. A plaintiff whose content claim is dismissed will replead as a product claim, and the answer to that is on the merits rather than under the statute.

Consider fee and sanctions exposure in jurisdictions with anti-SLAPP statutes, which can accelerate resolution and shift costs where the underlying claim targets protected activity.

And be accurate about what the platform did. Section 230 arguments that mischaracterise the platform's role are the ones that produce discovery and adverse findings.

Building the Record

The defence depends on facts about the platform that should exist before any complaint arrives.

A published content policy. Specific about what is prohibited and what enforcement follows. It supplies the standard for a (c)(2) argument and it is evidence that removals are content-based.

Terms of service that do not overpromise. Barnes v. Yahoo! exposure comes from specific commitments. Describe the process, reserve discretion, and avoid guarantees.

A moderation log. Report received, date, reviewer, policy provision applied, decision, and any appeal. This record is what makes moderation decisions defensible individually and collectively.

Notice and appeal mechanics where the terms promise them, applied consistently, because inconsistency is what turns a policy into a misrepresentation claim.

A repeat infringer policy implemented in fact, which is a condition of the 17 U.S.C. § 512 safe harbour and a discipline worth having independently.

A designated agent registration for DMCA purposes, kept current.

Accurate marketing. Verification claims, safety claims, and authenticity claims are the platform's own speech, and they should be reviewed by the same process as any other advertising claim. See Promotions and Advertising Compliance Toolkit.

Design documentation. Where a feature's safety was considered and mitigations were implemented, that record is the answer to a design-defect claim — which is a merits defence rather than an immunity one.

Escalation paths. Defined routes for law enforcement requests, court orders, and urgent safety reports, with records of what was done.

Retention. Moderation records retained long enough to cover the limitations periods for the claims that are actually brought.

The Marketplace Case

Marketplaces present the hardest version of the analysis, because they combine third-party listings with substantial platform activity.

Listing content is third-party content. Descriptions, images, and prices supplied by sellers are protected as to publishing claims.

Trademark claims are outside the statute. 47 U.S.C. § 230(e)(2) preserves intellectual property law, and there is no trademark equivalent of 17 U.S.C. § 512. Liability turns on contributory infringement — whether the marketplace continued to supply services to sellers it knew or had reason to know were infringing.

Which makes the notice-and-response programme the whole defence. A marketplace with a functioning brand registry, prompt removals, and a repeat-seller policy is in a substantially better position than one without.

Product liability claims are the live frontier. Whether a marketplace is a seller for product liability purposes is a state law question answered differently across jurisdictions, and it is largely independent of Section 230 because the duty alleged is not a publishing duty.

Fulfilment changes the analysis. A marketplace that warehouses, packages, and ships goods looks more like a seller than one that only connects buyers and sellers, and courts have drawn on those facts.

The platform's own representations. Verification badges, authenticity guarantees, and seller ratings generated by the platform are the platform's content.

Payment and escrow functions create their own obligations independent of content.

Consumer protection claims. State unfair and deceptive practices claims premised on the marketplace's own conduct — its representations, its fee disclosures, its dispute processes — are not publishing claims.

Practical programme. Brand registry enrolment, prompt takedown handling, seller verification proportionate to the risk, accurate marketing about what verification means, a repeat-infringer policy applied in fact, and a documented product safety process. See Running an E-Commerce Counterfeit Enforcement Program.

Regulatory Pressure

The statute has not changed materially, and the environment around it has.

State legislation. Statutes regulating platform design, age assurance, algorithmic transparency, and content moderation have proliferated, and their interaction with Section 230's preemption of inconsistent state law is being litigated.

Moody v. NetChoice addressed state laws restricting content moderation, confirming that curation implicates expressive interests while resolving the cases on the facial-challenge posture — which leaves the substantive questions to be worked out case by case.

Age-appropriate design codes impose duties on platform design rather than on content, which is deliberate: duties framed that way are not publishing duties. See Children's and Youth Privacy Toolkit.

Privacy statutes operate alongside and are not displaced. See State Privacy Compliance Toolkit.

Consumer protection enforcement targeting platform representations, dark patterns, and fee disclosures operates on the platform's own conduct.

Proposals to amend the statute are perennial and have not passed. Planning around a repeal is not advisable; planning for a landscape where state design regulation and consumer protection enforcement carry more weight than Section 230 litigation is.

The practical direction of travel. Immunity for third-party content remains robust. Liability for the platform's own design, representations, and processes is increasing. A compliance programme built around that distinction is aligned with where the exposure actually is.

Worked Example: The Complaint With Six Counts

A review platform is sued by a business that says false reviews destroyed it. Six counts.

Count one — defamation, based on user reviews. The platform is treated as the publisher of third-party content. Barred under 47 U.S.C. § 230(c)(1), and Zeran v. America Online disposes of the argument that notice changed the analysis.

Count two — defamation, based on the platform's summary rating. The rating is generated by the platform from third-party inputs. Courts have generally treated aggregation of third-party data as a neutral publishing function, but the argument is harder, and it depends on whether the platform contributed anything beyond arithmetic.

Count three — negligent design of the review system. Framed as a product claim after Lemmon v. Snap. Section 230 does not answer it; the merits do, and the design documentation becomes the defence.

Count four — promissory estoppel, based on the platform's stated policy of removing reviews found to be fake. Barnes v. Yahoo! supports the claim if a specific promise was made and relied on. The terms of service language decides it.

Count five — false advertising, based on the platform's claim that reviews are verified. The platform's own speech. Outside the statute entirely, and analysed under 15 U.S.C. § 1125(a)(1)(B) and state law.

Count six — trademark infringement, based on competitor advertising displayed against the plaintiff's name. 47 U.S.C. § 230(e)(2) preserves intellectual property law, so the analysis is ordinary trademark doctrine.

The outcome. Counts one and two dismissed on the pleadings. Counts three through six proceed, on their own merits, with no immunity available.

The lesson. The immunity is real and it is narrow in scope. It answers the content claim and nothing else, and a platform whose exposure is concentrated in its own representations and its own design gains very little from it.

Common Mistakes

Treating Section 230 as a general shield. It bars publishing claims about third-party content. It does not bar claims about the platform's own speech, its own design, or its own promises.

Overpromising in the terms of service. Specific removal commitments create Barnes v. Yahoo! exposure that the statute does not reach.

Marketing verification the platform does not perform. The platform's own representations are its own content and are directly actionable.

Requiring or structuring the unlawful element. Fair Housing Council v. Roommates.com turns on whether the tool required the content, and a form field that forces a prohibited category is the paradigm case.

Moderating for competitive reasons. Enigma Software Group USA v. Malwarebytes removes the (c)(2) protection where the motive is anticompetitive.

No moderation records. Without them, every removal decision is a swearing contest about motive.

Assuming the intellectual property carve-out is narrow. For trademark there is no safe harbour at all, and the marketplace counterfeit problem is a contributory infringement problem.

Ignoring the DMCA conditions. 17 U.S.C. § 512 requires a designated agent, a repeat infringer policy implemented in fact, and expeditious removal. A platform that assumes Section 230 covers copyright has no safe harbour at all.

Filing a Section 230 motion that mischaracterises the platform's role, which converts a clean dismissal into discovery.

Not moving early. The immunity's value is avoiding discovery, and a defence raised at summary judgment has already lost most of what it was worth.

The Plaintiff's Perspective

Firms act on both sides, and the plaintiff's analysis is the mirror image.

Do not plead the content claim as the lead count. It will be dismissed, and it colours the rest.

Identify what the platform did, not what users posted. The claim that survives is the one premised on the defendant's own conduct.

Look for representations. Verification claims, safety claims, screening claims, and authenticity guarantees are the platform's own speech and are directly actionable.

Look for promises. Terms of service commitments to remove, to notify, to investigate, or to appeal create Barnes v. Yahoo! exposure.

Look for design. A feature that causes harm independent of any particular content is a product claim under Lemmon v. Snap, and the duty alleged is the platform's own.

Look for off-platform knowledge. Doe v. Internet Brands permits a failure-to-warn claim where knowledge came from outside the publishing function.

Check the exceptions first. An intellectual property claim, a federal criminal predicate, or a FOSTA claim under 47 U.S.C. § 230(e)(5) is outside the statute and should be pleaded as such.

Plead the material contribution facts specifically. Where the platform's tools required or shaped the content, describe how — the drop-down, the mandatory field, the prompt.

Anticipate the motion. Draft each count so that its dismissal does not depend on facts outside the complaint, and so that the duty alleged is visibly independent of publishing.

And be realistic. Where the harm is genuinely third-party content and the platform merely hosted it, Zeran v. America Online still governs and the case is not viable.

Diligence Questions

What content does the platform host, and from whom?

Does the platform create, require, or structure any content that could be unlawful? The Roommates.com analysis applied to the actual product.

What does the platform say about itself? Verification, safety, screening, and authenticity claims, reviewed as advertising claims.

What do the terms of service promise? Specific commitments create contract and estoppel exposure.

Is there a published content policy, and is it applied consistently?

Are moderation decisions logged? With reviewer, policy provision, and outcome.

Is there a DMCA programme? Designated agent registered, repeat infringer policy implemented in fact, expeditious removal, and counter-notice handling under 17 U.S.C. § 512.

Is there a trademark notice programme? There is no safe harbour, so contributory infringement exposure turns on responsiveness.

Any pending claims, and how are they pleaded? Content claims and design claims have different trajectories.

Any FOSTA exposure, or state design or age-assurance obligations?

What design documentation exists for features that could support a product claim?

Questions Clients Ask

Are we protected if we know about the content? Yes, for publishing claims. Zeran v. America Online holds that notice does not defeat the immunity, which is the aspect most often misunderstood.

Are we protected if we edit content? Generally yes. Selecting, editing, and arranging third-party content are publisher functions. Creating content or requiring unlawful content is different.

Does our recommendation algorithm cost us the immunity? Force v. Facebook treats algorithmic arrangement as a publishing function, and that remains the prevailing view. Gonzalez v. Google did not resolve it, and it is the live question.

Can we remove content without liability? Generally yes, under (c)(1) and (c)(2). The exception is removal for anticompetitive reasons after Enigma Software Group USA v. Malwarebytes.

Does it cover trademark claims? No. 47 U.S.C. § 230(e)(2) preserves intellectual property law, and there is no trademark safe harbour.

Does it cover copyright? No, but 17 U.S.C. § 512 provides its own safe harbour if the conditions are met.

Does it cover state privacy claims? Generally the immunity applies to publishing claims, but privacy statutes imposing duties on the platform's own data practices are not publishing claims.

We promised in our terms to remove reported content. Is that a problem? Potentially. Barnes v. Yahoo! makes a specific promise enforceable independently of the statute.

Can we be sued for our design? Yes. Lemmon v. Snap permits a product claim, and the immunity does not answer it.

Do we need a DMCA agent? Yes, if you want the 17 U.S.C. § 512 safe harbour. Registration and currency are conditions.

Does the immunity apply to us if we are not a social network? The definition at 47 U.S.C. § 230(f) is broad and reaches marketplaces, forums, hosting providers, apps, and email services.

Will Section 230 be repealed? Proposals recur and none has passed. Build the programme around the exposure that is growing — the platform's own design, representations, and processes — rather than around a legislative forecast.

The One-Page Position

Platform liability position — [service], [date]. Content hosted: [categories], volume [N] items per period. Platform-generated content: [ratings / summaries / badges / editorial], reviewed as the platform's own speech on [date]. Product features assessed for Lemmon v. Snap design exposure: [N], with design documentation on [N]. Terms of service reviewed for Barnes v. Yahoo! commitments on [date]; [N] specific promises identified, [N] revised. Content policy published [date]; moderation log operating since [date]; [N] decisions recorded, [N] appeals. DMCA: agent registered [date], repeat infringer policy implemented, median removal time [N] hours. Trademark notice programme: [N] notices, median response [N] hours; no safe harbour available, contributory exposure assessed. Marketing claims about verification and safety reviewed [date]. FOSTA and state design obligations assessed [date]. Pending claims: [N], of which [N] are content claims and [N] are design or representation claims. Recommended actions: [revise terms commitment X / document design rationale for feature Y / correct verification marketing / build the moderation log].

Sector Notes

Social platforms. The core case. Publishing immunity is strong for user posts; the exposure has migrated to design claims about engagement mechanics, age assurance, and recommendation systems.

Marketplaces. Trademark and product liability dominate, and neither is answered by Section 230. The programme is a notice-and-response operation plus accurate representations about verification.

Review sites. User reviews are protected; platform-generated ratings and the site's own representations about verification are not. Terms commitments about removing fake reviews create the Barnes v. Yahoo! exposure.

Job and housing platforms. Fair Housing Council v. Roommates.com arose here, and any form field that requires or structures a protected-category preference is the paradigm material contribution.

Dating and social discovery. Failure-to-warn claims after Doe v. Internet Brands, design claims about safety features, and representations about screening.

Hosting and infrastructure. Broad immunity for hosted content, with the DMCA conditions under 17 U.S.C. § 512 doing the copyright work and the terms of service governing the customer relationship.

Gaming and user-generated worlds. User creations are third-party content; platform-provided tools raise the neutral-tools question; and virtual goods raise trademark issues outside the immunity. See Virtual Goods and Digital Brand Toolkit.

Messaging and communications. Content immunity applies; the Electronic Communications Privacy Act carve-out preserves privacy claims; and design and safety obligations under state statutes operate independently.

Anything with minors. Age-appropriate design obligations and children's privacy requirements are duties on the platform's own conduct and are unaffected by the immunity. See Children's and Youth Privacy Toolkit.

A Closing Note

The most useful thing to tell a platform client is that Section 230 protects them from the claim they are least likely to face and does nothing about the claims they are most likely to face.

Nobody with a real grievance sues a platform for hosting a post any more; the bar has read Zeran v. America Online. What arrives instead is a complaint about the platform's design, its promises, its representations, its screening, and its handling of reports — all of which are the platform's own conduct, all of which are outside the immunity, and all of which are governed by ordinary law that the company can actually manage.

Which makes the programme straightforward, if unglamorous. Do not promise what you will not do. Do not claim verification you do not perform. Document why features were designed as they were. Log the moderation decisions. Register the DMCA agent and run the repeat infringer policy. None of that is Section 230 practice, and all of it is what determines how a platform case actually goes.

The Analysis in Five Steps

For any complaint against a platform, this sequence resolves the position in an afternoon.

Step one — check the exceptions. Is any count an intellectual property claim, a federal criminal predicate, an Electronic Communications Privacy Act claim, or a FOSTA claim under 47 U.S.C. § 230(e)? Those are outside the statute and are analysed under their own frameworks.

Step two — for each remaining count, name the duty. Not the harm and not the content: the duty the plaintiff says was breached. Write it in one sentence per count.

Step three — ask whether that duty is a publishing duty. Deciding to publish, withdraw, postpone, or alter third-party content. If yes, 47 U.S.C. § 230(c)(1) bars the count and Zeran v. America Online answers the notice argument.

Step four — for counts that survive step three, ask whose content is at issue. Platform-generated content, platform representations, and content the platform required or structured under Fair Housing Council v. Roommates.com are the platform's own.

Step five — for counts about the platform's own conduct, litigate the merits. Design, promises, representations, and processes. The immunity is not the answer and the merits are.

Most complaints resolve into two or three counts dismissed at step three and three or four counts that proceed at step five. Presenting that division clearly — to the court and to the client — is more useful than arguing that the statute either does or does not apply to the case as a whole.

Working With Other Advisers

Litigation counsel for the motion strategy, the count-by-count division, and any anti-SLAPP or fee-shifting opportunity.

Trust and safety leadership, who own the content policy, the moderation workflow, and the logs that make the record. This is the function whose ordinary operations determine how a case goes.

Product and engineering, who own the design decisions that generate the claims Section 230 does not answer, and whose contemporaneous documentation of safety considerations is the defence.

Marketing, whose verification and safety claims are the platform's own speech and should be reviewed as advertising claims.

Privacy counsel, because the statutes imposing duties on data practices operate independently. See State Privacy Compliance Toolkit.

Brand protection, for the trademark notice programme that the intellectual property carve-out makes necessary.

Outside copyright counsel for the 17 U.S.C. § 512 programme, the agent registration, and the repeat infringer policy.

Regulatory counsel for state design, age assurance, and consumer protection obligations, which is where the direction of travel points.

What This Costs

The content policy and terms review. A week of counsel time, once, plus an annual refresh. The highest-return item in the list, because it removes contract exposure the statute does not touch.

The moderation log. Engineering work to instrument, then no marginal cost. It converts every removal decision from a swearing contest into a record.

The DMCA programme. Agent registration is nominal; the operational cost is the removal workflow, which most platforms need anyway.

The trademark notice programme. Real operational cost, scaling with volume, and unavoidable given the intellectual property carve-out.

Design documentation. Minutes per feature decision, captured in the ordinary product process rather than as a legal exercise.

A motion to dismiss. Modest against the cost of discovery, which is the whole point of asserting the immunity early.

Against all of that: a platform case that proceeds to discovery, with moderation records produced, product decisions deposed, and marketing claims examined — which is the outcome the programme is designed to avoid on the counts where avoidance is available.

A Suggested Reading Path

For the elements:

For the intellectual property carve-out:

For the contract layer:

Primary Authorities

| Authority | Proposition | |---|---| | 47 U.S.C. § 230 | Immunity; definitions; exceptions | | 17 U.S.C. § 512 | DMCA safe harbours | | 15 U.S.C. § 1125 | Lanham Act claims outside Section 230 | | 18 U.S.C. § 1591 | Sex trafficking; FOSTA reference | | 18 U.S.C. § 2421A | Promotion of prostitution | | Zeran v. America Online | Notice does not defeat immunity | | Fair Housing Council v. Roommates.com | Material contribution test | | Barnes v. Yahoo! | Promissory estoppel outside the statute | | Force v. Facebook | Algorithmic recommendation as publishing | | Gonzalez v. Google | Resolved on the underlying claim | | Twitter v. Taamneh | Aiding and abetting standard | | Doe v. Internet Brands | Failure to warn from off-platform knowledge | | Lemmon v. Snap | Product design claim outside the statute | | HomeAway.com v. City of Santa Monica | Duties on the platform's own conduct | | Enigma Software Group USA v. Malwarebytes | Good faith limit on (c)(2) | | Perfect 10 v. CCBill | IP exception construed as federal IP | | Moody v. NetChoice | Editorial curation and expressive interests |

Forms and Templates

There is no Section 230 form, because the defence is asserted in a motion rather than in a document a platform holds — but three artefacts determine whether the motion succeeds. The first is the terms of service, which should avoid promising what the platform will not consistently do, because Barnes v. Yahoo! holds that a specific promise to remove content can support a promissory estoppel claim that Section 230 does not bar. The second is the published content policy, which supplies the standard against which moderation decisions are measured and which makes a (c)(2) good faith argument available. The third is the moderation log, which records what was reported, when, what was decided, by whom, and against which policy provision — and which is the evidence that a removal was content-based rather than anticompetitive after Enigma Software Group USA v. Malwarebytes. The License Agreement Template is relevant where the platform licenses user content, and the Cease and Desist Template where the platform is enforcing rather than defending.

Related Toolkits and Checklists

The Online Terms and Consumer Contracts Toolkit covers the contract layer that sits alongside the statutory immunity and that supplies the claims Section 230 does not bar. The Platform Content Liability Checklist runs the analysis count by count. For the intellectual property exception, the Copyright Enforcement Toolkit covers the 17 U.S.C. § 512 regime that governs copyright claims against platforms, and the Anticounterfeiting and Border Enforcement Toolkit covers the marketplace trademark problem that the carve-out leaves outside the immunity.