What are the TTAB's rules and deadlines for submitting AI-generated or AI-assisted evidence in trademark proceedings?

The Direct Answer: How TTAB Deadlines Apply to AI Evidence

The Trademark Trial and Appeal Board (TTAB) does not maintain a separate, standalone rulebook for "AI evidence." Instead, AI-generated or AI-assisted materials submitted in inter partes proceedings—oppositions, cancellations, and concurrent use proceedings—are governed by the same deadlines and evidentiary frameworks that apply to all evidence: the Federal Rules of Evidence as applied by the Board, the TBMP (Trademark Board Manual of Procedure), the applicable scheduling order issued by the Board, and the notice-of-reliance requirements under 37 C.F.R. § 2.122. What has changed by mid-2026 is not a new set of deadlines but heightened scrutiny of how AI-assisted evidence is prepared, authenticated, and disclosed.

Also worth reading: What Are the Real Risks of AI-Generated Trademark Infringement for Brands in 2026? · What are the definitive trademark specimen requirements for 2026, especially regarding AI-generated content and digital goods? · What is the definitive trademark maintenance deadlines checklist for USPTO filings in 2026?

The core deadlines remain anchored to the accelerated case resolution schedule that the TTAB issues at the start of every proceeding. Under the standard schedule, each party typically receives a 30-day period for initial disclosures, followed by staggered 30-day discovery periods for each party, with testimony periods running 30 days per party thereafter. Evidence must be noticed for submission during the party's testimony period via a notice of reliance, and any rebuttal evidence must be filed within the rebuttal window—generally 15 days after the close of the opponent's testimony period in an opposition, subject to adjustment by the Board. If your evidence involves AI tools—whether machine-generated search reports, AI-summarized consumer surveys, synthetic mockups, or large-language-model outputs—the deadlines do not shift. You must identify that material in your disclosures and rely on it within the same windows as conventional evidence.

The practical consequence is that parties using AI have less room for error, not more. A late-disclosed AI-generated survey analysis cannot be cured by arguing that the tool was still "processing" data. The Board has consistently held that technological convenience does not extend deadlines, and motions to reopen testimony periods to admit newly generated AI analyses are routinely denied absent extraordinary circumstances under TBMP § 510.

Why AI Evidence Raises Distinct Deadline Risks

The reason AI materials deserve special attention is that they create authentication and disclosure burdens that can consume more of the fixed timeline than traditional evidence. Under Federal Rule of Evidence 901, a proponent must produce evidence sufficient to support a finding that the item is what it claims to be. For an AI-generated report, that means testimony or a declaration explaining the tool used, the inputs provided, the version of the model, and the methodology behind the output. Preparing that foundation takes time, and if you wait until the final week of your testimony period, you may find yourself unable to meet the notice-of-reliance deadline with a properly supported exhibit.

There is also the problem of accuracy. USPTO guidance issued in 2024 made clear that practitioners remain responsible for the accuracy of everything filed with the Office, including content generated or refined by AI tools. That principle carries directly into TTAB practice: an AI-hallucinated citation, a fabricated survey statistic, or a misattributed third-party article discovered late in the process forces a scramble—withdrawal of the exhibit, potential sanctions exposure under 37 C.F.R. § 11.18, and damage to credibility before the Board. Because the correction window is constrained by the same fixed schedule, verification of AI output should begin the day the material is generated, not the day it is due.

Finally, discovery obligations interact with AI in ways that affect timing. If a party uses AI tools to analyze documents produced in discovery, the opposing party may propound interrogatories or requests for admission about the methodology. Responses are due within 30 days under the standard rules, and evasive answers invite motions to compel that eat into remaining testimony periods. Parties who disclose their AI workflows early—in initial disclosures or in response to reasonable interrogatories—tend to avoid these timing collisions entirely.

Practical Steps: Meeting Every Deadline When Using AI Tools

The first step is calendar discipline built around the Board's scheduling order. Mark three dates immediately when the order issues: the close of your discovery period, the opening date of your testimony period, and the notice-of-reliance deadline, which falls at least 15 days before the close of your testimony period under standard practice. Work backward from those dates. If you intend to submit an AI-assisted consumer survey, the underlying survey fieldwork must be completed well before the testimony period opens, because the AI analysis layer adds days for validation, expert review, and declaration drafting.

Second, build a verification log for every AI-generated exhibit. Record the tool name and version, the date of generation, the prompts or inputs used, and the human review performed. This log serves double duty: it supports authentication under Rule 901 and demonstrates diligence if the Board questions the material's provenance. Practitioners using AI-assisted prior-use searches or marketplace screenshots enhanced by AI upscaling should preserve the original, unmodified source files alongside the processed versions, since the originals are what opposing parties will demand in discovery.

Third, use the disclosure mechanisms proactively. Initial disclosures under the Board's procedures require identification of witnesses and key evidence early in the case. Identifying an AI-generated survey or analytics report at that stage gives the opposing party time to inspect it and gives you buffer against motions to exclude based on surprise. Fourth, file notices of reliance precisely: list each AI-derived exhibit, attach it where required, and confirm service dates. A defective notice of reliance is one of the most common—and most avoidable—procedural failures in TTAB cases involving novel evidence types.

Comparison: Traditional Evidence vs. AI-Assisted Evidence Under TTAB Rules

The table illustrates the central point: the procedural architecture is unchanged, but the evidentiary overhead of AI materials is higher. Budgeting that overhead into the fixed timeline is the difference between admissible evidence and a waived exhibit.

Common Mistakes That Cost Parties Their Evidence

The most frequent error is treating AI output as self-authenticating. A screenshot of a chatbot response, an automated similarity score from an AI clearance tool, or a machine-generated market-share chart is not evidence until a competent witness adopts it and explains its basis. Parties who attach such materials without declarations frequently see them struck or given no weight, effectively losing the benefit of the exhibit despite meeting the filing deadline.

A second mistake is missing the notice-of-reliance window because the AI analysis was not finished. The Board will not hold a testimony period open for a pending algorithmic run. If the analysis is not ready, the choice is to proceed without it or seek consent from the other side—an outcome that becomes unlikely once the adversary realizes the evidence could hurt them. Third, some parties attempt to introduce AI-generated "consumer perception" narratives without any underlying survey, which the Board treats as argument rather than evidence. Fourth, practitioners sometimes fail to correct AI-fabricated citations promptly upon discovery, converting an honest error into a candor problem. Prompt withdrawal and notification protect both the record and the practitioner's standing. Fifth, over-reliance on AI-drafted briefs near filing deadlines has produced visible errors in recent years; the Board expects signed filings to reflect attorney review, and the signature requirement under 37 C.F.R. § 2.11(a) makes that responsibility explicit.

When to Act: Timing Your AI Evidence Strategy From Day One

Decisions about AI evidence belong in the case-strategy phase, not the eleventh hour. At the outset of an opposition or cancellation, evaluate which elements of proof—likelihood of confusion surveys, fame evidence, marketplace usage, dilution metrics—could benefit from AI-assisted analysis, then assess whether the resulting exhibits can be completed, validated, and disclosed inside the standard schedule. Fame cases illustrate the stakes: proving fame requires extensive third-party usage and sales evidence, and AI tools can accelerate document review, but the resulting compilations still need declarant support and must be noticed within the testimony period.

If the standard schedule genuinely cannot accommodate necessary AI-dependent work—for example, a full survey requiring several months of fieldwork—the proper vehicle is a motion to modify the scheduling order under TBMP § 503, filed as early as possible with a showing of good cause. Motions filed on the eve of a deadline rarely succeed. Conversely, if the AI work is complete early, consider stipulating with the other side to advance the testimony period; the Board generally welcomes efficiency, and early submission removes the risk of last-minute technical failures in ESTTA, the Board's electronic filing system, which experiences predictable congestion near major deadlines.

Costs, Resources, and Where AI Actually Saves Money

Using AI in TTAB evidence preparation shifts costs rather than eliminating them. Commercial survey platforms and AI-assisted analytics subscriptions typically run from modest monthly fees into the low thousands annually, while a traditional likelihood-of-confusion survey from a recognized vendor commonly costs tens of thousands of dollars. AI-assisted document review can compress associate hours substantially in fame-heavy records, though expert declaration fees—often several thousand dollars per declaration—remain unavoidable because the Board weighs expert foundations heavily. The genuine savings come from speed: faster review means fewer rushed filings, fewer extension requests, and lower risk of the wasted expense of preparing evidence that misses its window. Parties should also account for the cost of verification labor, which is non-negotiable; skipping it risks sanctions and rework that dwarf the original savings.

The Broader Regulatory Context Heading Into Late 2026

TTAB practice sits inside a fast-moving regulatory environment. India's proposed rules on synthetic media, reported by World Trademark Review, would impose new due diligence obligations on intermediaries handling AI-generated content—a signal that jurisdictions worldwide are tightening accountability for synthetic materials. In the United States, the USPTO's existing guidance on AI use in filings, combined with the Executive Branch's evolving AI policy instruments, points toward greater documentation expectations rather than looser ones. For TTAB practitioners, the trajectory is clear: the deadlines will stay fixed, but the evidentiary hygiene expected around AI-generated material will keep rising. Building disciplined verification and disclosure habits now positions parties to meet whatever formalized requirements emerge next, without scrambling against a 30-day clock.