copyright in model codes after Georgia v. PublicResource.Org

Older case, but worth working through!

International Code Council, Inc. v. UpCodes, Inc., 2020 WL 2750636 (S.D.N.Y. May 26, 2020)

ICC claimed that defendants infringed its copyrights in forty model building codes (I-Codes) by posting them and derivative works on their website, UpCodes. The court endorsed the basic proposition that there can be no liability for sharing the law with the public, whether as a matter of copyrightability or fair use, but found that some of UpCodes’ “redlines” might go too far and thus denied cross-motions for summary judgment. It’s a great survey of issues that remain after the Georgia v. Publicresource.org case and well worth reading—especially on merger.

State/local adoption of model codes benefits everyone involved, including the public, by aligning the law with industry best practices. ICC was founded in 1994 by three regional standards development organizations (SDOs) that developed similar codes, who donated their copyrights in model codes to ICC in 2003 “and envisioned that ICC would produce coordinated national building codes. Its codes have been adopted into law by jurisdictions across all fifty states.

One of the “foremost” reasons ICC develops model codes is to get them enacted into law. It has a dedicated government relations department that helps jurisdictions enact its codes. The development process “allows for public participation and comment at no cost,” “involves significant participation by government representatives,” and includes revisions every three years.  “ICC currently incurs the up-front costs of its code development process and recoups the relevant costs at least in part from the sale and licensing of its copyrighted codes.” ICC charges “modest” prices compared to those for other technical reference works and donates hundreds of copies of its model codes to libraries and other jurisdictions throughout the US. Members of the public can purchase model codes and guidelines online. ICC also sells a variety of supplemental materials and services, such as code commentary, study companions, handbooks, user’s guides, and training materials incorporating model code text. And:

ICC also makes its model codes, and some government codes that adopt the model codes, available for free on its website in a digital library called publicACCESS. The publicACCESS reading room makes printing, copying, and downloading of the codes difficult, though it does not necessarily completely prevent such functions. ICC publicizes that the free access it provides to its model codes and the enacted state and local codes is on a read-only basis. ICC relatedly offers a paid service called premiumACCESS, which provides users with access to additional features such as full access to commentaries, and tools to highlight, bookmark, and annotate the codes.

Meanwhile, UpCodes is a startup providing “easy and convenient access to materials of particular importance to members of the architecture, engineering, and construction (‘AEC’) industries, such as the state and local building codes that governed their projects.” Like publicACCESS, it has some free services and others that are available only to paid users. There are two versions of relevance to the suit: pre-suit UpCodes made the forty I-Codes at issue freely available to anyone to view, print, copy, and download, but redirected users to ICC’s website for at least one (the International Zoning Code 2015). The forty ICC model codes were posted in a section titled “General Building Codes” and identified by their model code names, but also identified jurisdictions that had adopted the model codes on the same page, at least as to some of the codes. Meanwhile, the paid access portion displayed governmentally enacted building codes that adopted ICC’s model codes with amendments. “[A]dditions by the enacting state or local jurisdiction were displayed in green, while all model code text that a jurisdiction did not adopt was struck-through in red, much like in a redline.”

Current UpCodes “purports to post only enacted state and local building codes, rather than any model codes as such,” and doesn’t charge for access to enacted laws. It also now shows only the titles or headings of the deleted model code provisions in struck-through red text, rather than all portions of the model codes that were not adopted. But it still uses the “trademarked” names of ICC’s model codes at various points, e.g. by noting on a state/local code page that the enacted code adopts a particular model code with or without amendment.

There is an extensive and nuanced review of the case law, but the bottom line is simple: “the I-Codes as Adopted are in the public domain, because they are in fact enacted state and local laws binding on the enacting jurisdictions’ constituents.” (The court declined to rely on collateral estoppel against ICC’s predecessors in interest.)  However, it is possible that defendants infringed by posting the codes “as model codes or the I-Code Redlines … [T]he record is ambiguous as to whether what the Defendants actually post constitutes ‘the law’ alone.”

The court rejected ICC’s arguments that, if adoption of model codes prevents enforcement of copyright in those codes, that’s an unconstitutional taking, and its related argument that 17 U.S.C. § 201(e) prohibits government expropriation of copyrights.

Takings: “Private parties do not have reasonable investment-backed expectations in property or information voluntarily provided to government, beyond what is explicitly provided by the government itself.” Given that ICC urged government adoption of its codes, it could not complain when that happened, even if it wanted to have its copyright and enact it too:

Even if ICC did not expect that its encouragement of government adoption would prevent it from enforcing its copyrights as to the I-Codes as Adopted, that legal consequence flows from the federal law of the public domain rather than from unjust action by the state or local jurisdictions. Far from coercing ICC to give up its copyrights, the jurisdictions are following ICC’s advice that the I-Codes would protect their citizens better than would the jurisdictions’ trying to draft complex technical codes from scratch.

Nor did it matter that “many jurisdictions sign licensing agreements respecting ICC’s copyrights.” ICC retained copyrights in its model codes as model codes, but that didn’t answer the question of whether the model codes became “laws” governing the public. Regardless, those agreements could not override the public’s right to “freely share the laws that govern them.”

As for §201(e), that applies only to individual authors, and also more fundamentally “addresses government actions avowedly intended to coerce a copyright holder to part with his copyright, so that the government itself may exercise ownership of the rights.”

How broad is this holding?

A privately-authored work may “become the law” upon substantial government adoption in limited circumstances, based on considerations including (1) whether the private author intended or encouraged the work’s adoption into law; (2) whether the work comprehensively governs public conduct, such that it resembles a “law of general applicability”; (3) whether the work expressly regulates a broad area of private endeavor; (4) whether the work provides penalties or sanctions for violation of its contents; and (5) whether the alleged infringer has published and identified the work as part of the law, rather than the copyrighted material underlying the law. These considerations may not all be strictly necessary or exhaustive, but are guideposts to assess whether notice of the purported copyrighted work is needed for a person to have notice of “the law,” such that due process concerns would effectively categorically outweigh the private author’s need for economic incentives.

However, “SDOs may still sue for infringement if a defendant copies their model codes as model codes or indiscriminately mingles the enacted portions of the model codes with portions not so enacted.”

Though there were some factual issues remaining, the framework was clearly highly defense-favorable. Even the trademark-y, “others are free to copy the law, they are not free to copy the copy” bit favored defendants: “While it would likely be inappropriate for a user to post the model codes without any indication that they have been adopted into law, the Court is not persuaded that it would be improper to identify in such posting both an enacted law and where that law derived from. Whether a state or local jurisdiction has referenced a private work is a matter of fact, and is not equivalent to posting the private work itself.”

A jury could find that posting model codes as model codes or “indiscriminately” mingling enacted text with nonadopted text constituted willful infringement, “particularly considering that Historic UpCodes also hid state and local amendments to the I-Codes behind a paywall.” (A footnote also commented that redlining isn’t really indiscriminate because of its well-understood signalling function.) But a member of the public can post enacted laws “and state the simple fact that those laws are derived from the I-Codes.” 

One interesting argument: UpCodes’ posting of the law was underinclusive, “because they do not post other statutory provisions that might affect the enacted model code text,” including definitional statutes. The court disagreed; for example, South Holland’s enacting ordinance states that “the International Building Code (2012) be and is hereby adopted as the building code of the Village of South Holland in the State of Illinois.” “While South Holland undoubtedly has other laws that pertain to building safety, the Court cannot conclude that Defendants’ description of the IBC 2012 as the building code of South Holland was inaccurate when South Holland’s ordinance states the same.” It was true that defendants needed to post amendments to the I-Codes themselves “to accurately portray the enacted law,” it wasn’t practical to require them to post any “additional legal material that may bear on the enacted model text’s full meaning,” since such interactions could go on infinitely.

The court denied the defendants’ motion for summary judgment because there were genuine factual disputes suggesting that at least some codes posted on Current UpCodes “indiscriminately mingled” enacted text with unadopted model text, e.g. posting a code as part of Wyoming’s building code with all the appendices, when Wyoming didn’t adopt any appendices (including one focusing on tsunami-generated flood hazards). 

As to the redlines, they were not in the public domain just because they “reflect[ ] the work of lawmakers just as much as the enacted text of the law.” That had to be assessed as fair use.

Merger: Historic UpCodes displayed the I-Codes as model codes in full, including even their copyright pages. The parties disputed at what time merger should be assessed: at the creation stage (in which case the ideas are separable from the text) or the putative infringement stage (in which case they aren’t separable from the law as enacted). While the Federal Circuit has said that merger should be assessed at the time of copying (Google v. Oracle), the court here disagreed, elegantly pointing out that this contradicted the Federal Circuit’s own framing of the Second Circuit approach, which is to assess merger as a matter of infringement, not copyrightability:

[I]f that defense can turn only on considerations that relate to the time of initial copyrightability, rather than the time of the activity that may expose a defendant to liability, it is hard to see how courts are meaningfully considering merger “in the context of alleged infringement” and under the rubric of substantial similarity.  Far from providing courts with “a more detailed and realistic basis for evaluating” defendants’ merger claims, such reasoning requires courts to ignore potentially relevant evidence from defendants while shifting a burden of proof to them.

Thus, merger requires assessing “whether previously copyrighted language has become essential to the expression of, or integrated with, a legal conception.” The enacted law is a fact. “Even though the model codes themselves have not become the law, one would need to use those model codes’ precise language to express laws that had adopted the codes by reference in their entirety.”

What about the copyright pages? “[D]e minimis under the circumstances.” So too with the current UpCodes approach of showing model code headings struck-through in red when displaying enacted codes that amend the model codes, “bearing in mind the headings’ minimal contribution to the overall model codes.”

Again, there were factual issues about what exactly defendants posted that prevented summary judgment for them:

There are some pages on Historic UpCodes where Defendants appear to have posted model building codes without mentioning adopting jurisdictions. On still other pages, the model code is accompanied by a section listing states that adopted the model code, but it is unclear whether those states adopted the code without amendment (such that the model code was the only way to express their laws, even if not explicitly identified as such) or whether those states amended the model codes (such that the model code text was not the only way to express those laws). That Defendants made government amendments to the I-codes available only to paying users may also cut against the notion that they meant only to express the law.

Fair use: The court would have reached the same result on the I-Codes as adopted if it had relied on fair use, and the redlines might be fair use. Purpose: “Posting enacted laws for the purpose of educating members of the public as to their legal obligations may be transformative, even if the enacted laws are identical to other copyrighted works.” And commerciality isn’t very important when the use is educational/informative/transformative.

Whether the redlines were transformative was a closer call. “In a sense, the I-Code Redlines are like legislative history showing what the state and local jurisdictions explicitly decided to add or delete when adopting the model codes.” The PRO case expressed concern that a state might try to “monetize its entire suite of legislative history.” But those were works authored by judges and legislatures, and the redlines might be “less probative” than legislative history or annotations. “As ICC points out, knowing that Wyoming did not adopt an appendix on tsunami-generated flood hazards probably does not help Wyoming residents comprehend their legal duties.” And that the redlines were only available to paying customers could partially offset transformativeness.

Nature of the work: highly factual, even as to the redlines; favored defendants.

Amount copied: As in the DC Circuit’s ASTM case, where a defendant “limits its copying to only what is required to fairly describe the standard’s legal import, this factor would weigh strongly in favor of finding fair use here, especially given that precision is ten-tenths of the law.” However, that likely weighed against fair use for the redlines.

Market effect: Again relying on ASTM, the court considered (1) the marginal harm caused, given that SDOs make their standards available for free in controlled reading rooms without hurting sales of their standards; (2) the harm caused by making available only the portions of the standards that “became law” versus the markets for the complete standards; (3) the market for derivative works, “particularly considering that the standards are regularly updated and that the private parties most interested in the standards would presumably remain interested in having the most up-to-date ones.” ASTM also noted that ICC remains profitable, despite its predecessors in interest having lost similar cases, through sales of copies of the I-Codes and other services such as consulting, certification, and training.

This factor, unlike the others, could potentially weigh against copying the I-Codes as adopted, because if they were adopted without amendment they were effective substitutes for the model codes. There was also a genuine dispute on the effect on the market for derivative works, though the court doubted that any dispute was material given the combined weight of the other three factors. The effect of the redlines was also unclear; they could have competed with ICC’s own redlines, but “the record tends to focus on a variety of derivative works that are not redlines, such as training and certification documents, user’s guides, handbooks, and code commentary.” (In a footnote, the court noted that UpCodes could substitute for the controlled reading rooms/online libraries—but, I note, if those are free to users, then it’s not clear what “market” ICC loses from that.)

For the same reasons, the court declined to rule on whether defendants’ infringement (if it existed) was willful. A reasonable jury could so find:

Defendants undisputedly copied the I-Codes without ICC’s authorization, and ICC cites multiple statements suggesting Defendants knew doing so would displease ICC and possibly harm ICC’s business. ICC also argues that Defendants were at least reckless insofar as they posted the I-Codes without seeking the advice of counsel.

But they might nevertheless have believed their actions were entirely legal “based on their understanding of the law,” as their public statements indicated, and the jury might believe that they couldn’t afford to hire a lawyer; they might not have recklessly disregarded the possibility of infringement. (If the district judge can’t figure out whether there was infringement even after extensive briefing and analysis, could it really be reckless to think there wasn’t?)