Apple's $250M Siri settlement: the price of an over-promised Apple Intelligence, in the public record

The problem

When Apple introduced the iPhone 16 line and the iPhone 15 Pro and Pro Max, it marketed the phones as a breakthrough in on-device generative AI under the Apple Intelligence brand, with a substantially upgraded Siri as the centerpiece [source]. A consumer class action, Landsheft v. Apple Inc., brought by named plaintiff Peter Landsheft in the U.S. District Court for the Northern District of California (No. 5:25-cv-02668-NW), alleged those AI features were over-promised: according to the complaint, Apple advertised “AI capabilities that did not exist at the time, do not exist now, and will not exist for two or more years” [source]. Reporting the settlement, TechCrunch summarized the core allegation as a claim that “Apple exaggerated the breadth of features Apple Intelligence would bring” ([source]). The operative filing itself, the Second Consolidated Amended Class Action Complaint (ECF No. 76, filed May 1, 2026), puts the allegation in the plaintiffs’ own words: after buying an iPhone 16 and setting it up, a named plaintiff found the phone “did not have the promised features and is substantially inferior to what Apple advertised” [source]. The complaint further alleged Apple’s advertisements “saturated the internet, television, and other airwaves to cultivate a clear and reasonable consumer expectation that these transformative features would be available upon the iPhone’s release” ([source]).

What was built

The disputed product is Apple Intelligence, Apple’s on-device and Private-Cloud generative-AI system, and specifically the enhanced Siri assistant marketed alongside the iPhone 16 launch as able to act with awareness across a user’s apps [source]. The resolution being examined here is not a product but a settlement: a $250 million, non-reversionary cash fund resolving the class claims, negotiated by class counsel Cotchett, Pitre & McCarthy, LLP ([source]). The court preliminarily approved that settlement “as fair, reasonable, and adequate to the Settlement Class, as falling within the range of possible final approval,” after holding a hearing on the motion on July 8, 2026 [source].

The outcome

Apple agreed to pay $250 million to settle the case; as AppleInsider reported, “the parties settled for $250 million, offering U.S. Settlement Class Members $25 per eligible device” [source]. The Next Web put the same figure firsthand: “Apple agreed to pay $250m to settle a US federal class action over the marketing of Apple Intelligence-powered Siri features” ([source]). The base payment is $25 per eligible device, rising to as much as $95 per device if the number of valid claims is low [source], a schedule the court’s own order states directly: claimants are entitled to “a presumptive payment from the Net Settlement Amount of $25 per Eligible Device,” with “a per-device cap of $95” ([source]). The class is large: the settlement “will pay between $25 and $95 per eligible device to roughly 36 million US iPhone 16 and iPhone 15 Pro owners who purchased between 10 June 2024 and 29 March 2025” [source], and that purchase window of June 10, 2024 to March 29, 2025 is fixed in the court’s own order ([source]). The approval so far is preliminary, not final: Apple’s “settlement proposal received signed preliminary approval from a judge last week after a hearing that occurred on July 8,” and “the final approval hearing for payouts (not just the terms of the deal) isn’t until September 29, 2027” [source]. The court’s order confirms that date as the “Final Approval Hearing to determine final approval of the Agreement on September 29, 2027” ([source]). These are allegations resolved by settlement, not an adjudicated finding that Apple deceived consumers.

A note on sourcing

The procedural spine of this account (that the case exists, who the class is, the purchase window, class counsel, preliminary approval, and the 2027 final-approval hearing) rests on the court’s own preliminary-approval order (Source 1, Tier 1). Two of the substantive claims are now anchored to primary filings as well: the $25 base and $95 per-device cap are stated verbatim in that order (which quotes the Settlement Agreement’s schedule), and the over-promise allegation is quoted from the operative Second Consolidated Amended Complaint itself (Source 6, Tier 1). The weakest load-bearing figure that remains is the $250 million gross fund total and the ~36 million device count: neither is restated in the order, so both rest on independent tech-press reports (Sources 2 to 5, Tier 2). Four mutually independent outlets agree on those numbers, but the gross fund total lives in the Settlement Agreement, not in any document reproduced here.

Related case files

• FTC v. Workado, an AI content detector advertised at 98% accuracy that independent testing put at 53%: the same gap between a marketed AI capability and the delivered one, reached through a regulator rather than a class action.
• SEC AI-washing action against Rimar Capital: $310,000 in civil penalties for overstated AI: AI-washing named as such by a federal regulator, the enforcement counterpart to this private redress.
• FTC v. DoNotPay, the “world’s first robot lawyer” and $193,000 over deceptive AI claims: another AI product whose marketed abilities outran what it could do.
• SEC AI-washing order: Presto’s drive-thru “voice AI” needed a human on every order: the same gap between a marketed AI capability and the delivered one, here in operations rather than consumer marketing, reached by a regulator.