Somewhere right now, a person is finishing a sentence. The moment the words land on the screen—saved, autosaved, captured in a way that lasts more than a fleeting instant—that person becomes the owner of a federal property right that will outlive them by seventy years. No form was filed. No fee was paid. No lawyer was consulted. No notice was stamped on the page. Copyright simply attached, silently and automatically, the way condensation forms on a cold glass. This is one of the strangest and most powerful features of American intellectual property law: the most valuable creative rights in the United States are acquired by accident, by millions of people, every single day.

That automatic quality is also where most of the confusion starts. Because copyright costs nothing to obtain, people assume it must protect everything—your business idea, your recipe, your brand name, your clever method, the facts you painstakingly gathered. It does not. And because copyright is invisible at the moment of creation, people assume the law is loose about what they can do with other people's work—copy a photo here, sample a beat there, "it's only fair use." It is not, or at least not nearly as often as people hope. Copyright is at once broader and narrower than intuition suggests, and the gap between intuition and doctrine is where lawsuits live.

This article is the hub. It is a guided tour of the entire copyright system: what the law protects and what it refuses to protect, who owns it, how long it lasts, how you transfer it (and how authors can claw it back decades later), what counts as infringement, what fair use really means after two recent Supreme Court decisions reshaped it, how registration works and why the timing matters enormously, what the Digital Millennium Copyright Act added to the picture, and how any of this works once you cross a border. Along the way we will meet the cases every copyright lawyer carries in their head—Feist, Baker v. Selden, Google v. Oracle, Warhol v. Goldsmith, Fourth Estate—and a few invented companies (say hello to Acme Corp. and Bluebird Records) to make the abstractions concrete. Where you want to go deeper, we link to companion articles, such as the step-by-step copyright registration guide and the plain-answer copyright FAQ. Treat this as the map; the linked articles are the closer surveys of particular territory.

The Foundation: Where Copyright Comes From

Copyright in the United States is not a common-law accident that judges invented. It is written into the Constitution itself. Article I, Section 8, Clause 8—often called the Intellectual Property Clause or the Copyright Clause—empowers Congress "To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries." Read it carefully, because every word does work. "To promote the Progress" tells us the system has a purpose beyond rewarding authors: it is a utilitarian bargain meant to enrich the public storehouse of knowledge and culture. "For limited Times" tells us copyright must eventually end—there is no perpetual copyright in America, however long the term has grown. And "Authors" and "Writings" tell us the right belongs to creators of expressive works, a category the Supreme Court has read broadly but not infinitely.

Congress exercised that power most recently and comprehensively in the Copyright Act of 1976, codified in Title 17 of the United States Code. The 1976 Act is the statute you will see cited throughout this article by section number (Section 102 for subject matter, Section 106 for exclusive rights, Section 107 for fair use, and so on). It took effect on January 1, 1978, and it replaced an older 1909 regime that had governed for nearly seventy years. That date—January 1, 1978—is a hinge in copyright law. Works created on or after that date live under the modern rules; works from before it can be governed by older rules about notice, publication, and renewal that still trip up the unwary. We will return to that seam when we discuss duration.

Two institutions administer the system. The United States Copyright Office, a department within the Library of Congress, registers claims, records transfers, issues regulations (found in Title 37 of the Code of Federal Regulations), and publishes the Compendium of U.S. Copyright Office Practices, a thick manual that is the closest thing copyright has to an operating handbook. The federal courts—copyright is exclusively federal jurisdiction; there is no such thing as a state copyright claim for works within the Act's scope—decide infringement suits and ultimately say what the statute means. Keep both in mind: the Copyright Office tells you how to register and what it will and will not accept; the courts tell you what your rights are worth when someone tramples them.

What Copyright Protects: Section 102 and the Magic Words

Section 102(a) of the Copyright Act is the gateway. It extends protection to "original works of authorship fixed in any tangible medium of expression." Pack that phrase into a suitcase and you find three requirements, each of which has spawned a small library of case law: the work must be original, it must be a work of authorship, and it must be fixed.

Fixation is the easiest to grasp and the one people forget. A work is "fixed" when it is captured—by the author or under the author's authority—in a medium stable enough that the work can be perceived, reproduced, or communicated for more than a transitory duration. Ink on paper, pixels on a saved file, a recorded song, a sculpted block of marble, a building as constructed: all fixed. A live jazz improvisation that no one records is not fixed, and so it is not protected by federal copyright—a startling result that means a musician who hears a brilliant unrecorded solo and reproduces it commits no copyright infringement (though states may offer some protection, and recording the solo would create a fixed work). A choreographer's dance is protected only once it is notated or filmed. An off-the-cuff speech enjoys no federal copyright until it is written down or captured. Fixation is the price of admission, and it is usually paid the instant a creator hits "save."

Authorship carries a quiet but increasingly important limitation: the author must be human. The Copyright Office and the courts have held that works produced by nature, by animals, or by machines without sufficient human creative control are not copyrightable. The famous "monkey selfie" litigation—where a crested macaque named Naruto pressed a camera shutter and a photographer claimed the resulting image—ended with courts confirming that a monkey cannot hold a copyright. More consequentially for 2026, the Copyright Office has taken the position, upheld in Thaler v. Perlmutter, that purely AI-generated images with no human authorship cannot be registered, and that a human must contribute the original, creative expression for a work involving generative AI to qualify. The line between "the human directed the machine enough" and "the machine did it" is one of the most actively litigated and regulated questions in copyright today; we explore it in depth in our companion piece on AI-generated inventions and in the broader overview of AI legal issues.

Originality is the deepest of the three, and it is the subject of the single most important copyright case most non-lawyers have never heard of: Feist Publications, Inc. v. Rural Telephone Service Co., 499 U.S. 340 (1991). Rural Telephone, a Kansas phone company, published a standard white-pages directory—names, towns, and numbers listed alphabetically. Feist, a competitor compiling a regional directory, copied thousands of Rural's listings. Rural sued for copyright infringement. The intuitive answer is that Rural worked hard to gather all those numbers, so surely it owns them. The Supreme Court, unanimously, said no. Justice O'Connor's opinion demolished the so-called "sweat of the brow" theory and held that originality—not effort—is "the touchstone of copyright protection." A garden-variety alphabetical phone book lacks the "minimal degree of creativity" the Constitution requires, because facts are not original to anyone (they are discovered, not authored), and an obvious, mechanical arrangement of facts adds nothing. Feist teaches the two pillars of originality: the work must be independently created (not copied from another) and must possess at least a modicum of creativity. The bar is famously low—a "minimal" spark suffices, and the vast majority of works clear it—but it is not zero. Pure facts, and the garden-variety compilation of them, fall through.

Section 102(a) also gives a non-exhaustive list of categories of protectable works, and it is worth knowing because the Copyright Office's registration system tracks it. The categories are: literary works (which include computer programs and most compilations); musical works, including any accompanying words; dramatic works, including accompanying music; pantomimes and choreographic works; pictorial, graphic, and sculptural works (everything from fine art to maps to technical drawings); motion pictures and other audiovisual works; sound recordings (the fixation of a performance, legally distinct from the underlying musical composition—a distinction that bedevils the music industry and which we unpack in music licensing in the streaming era); and architectural works (added in 1990, protecting building designs as built or in plans). These categories are meant to be read broadly. The phrase "including but not limited to" precedes them, so a work that does not fit neatly can still qualify.

What Copyright Does Not Protect: The Idea-Expression Line and Baker v. Selden

If Section 102(a) is the gateway, Section 102(b) is the wall. It declares that copyright protection extends to no "idea, procedure, process, system, method of operation, concept, principle, or discovery, regardless of the form in which it is described, explained, illustrated, or embodied." This is the idea-expression dichotomy, and it is the most important limiting principle in all of copyright. Copyright protects the particular expression of an idea—the specific words, notes, brushstrokes, or lines of code—but never the underlying idea, system, or method itself. The idea belongs to everyone; only the author's expression of it is owned.

The dichotomy traces to a wonderful old case that every law student reads: Baker v. Selden, 101 U.S. 99 (1879). Charles Selden wrote a book explaining a novel system of bookkeeping, and the book included blank ledger forms with ruled columns and headings designed to implement his system. Baker began selling books with similar forms. Selden's widow sued, claiming Baker had infringed the copyright in Selden's book. The Supreme Court drew the line that still governs: Selden's book—his explanation, in his words, of how the bookkeeping system worked—was protected by copyright, and Baker could not copy that text. But the system itself, the method of bookkeeping, and the blank forms necessary to use it, were not protected by copyright at all. If Selden wanted a monopoly on the method, the Court said, his remedy was a patent, not a copyright. Copyright "cannot give to the author an exclusive right to the methods of operation which he propounds."

Baker v. Selden echoes through modern law in surprising places. It is why a recipe's list of ingredients and basic steps is not copyrightable (it is a functional method), though the headnotes, stories, and creative descriptions around it might be. It is why rules of a game are free for all to use even though a particular rulebook's prose is protected. And it is why, in Google LLC v. Oracle America, Inc., 593 U.S. 1 (2021), the functional, organizing structure of a software interface sat uneasily at the edge of copyrightability—a question the Court ultimately sidestepped by deciding the case on fair use, as we will see. The doctrine has a sharper cousin called merger: when there is only one or a very small number of ways to express an idea, the idea and expression are said to "merge," and copyright will not protect that expression either, lest the law hand someone a monopoly on the idea by the back door. If there is essentially one efficient way to write a particular short snippet of code or to phrase a simple factual statement, copying it does not infringe.

The statute and the Copyright Office add a list of things that fall outside protection for related reasons. Titles, names, short phrases, and slogans are not copyrightable—too short to carry original authorship (though they may be protectable as trademarks, which is one of many places copyright and trademark diverge; see copyright vs. trademark vs. patent). Familiar symbols and designs, mere typeface variations, and simple listings (like ingredient lists) are out. Blank forms that merely record information, following Baker, are out. And facts and ideas, per Feist and Section 102(b), are always free. A useful mental model: copyright is a fence around a creator's particular way of saying or showing something. Stand outside the fence and you are free to express the same idea your own way; the only thing forbidden is copying what is inside.

The Bundle of Exclusive Rights: Section 106

Owning a copyright does not mean owning a single right. It means owning a bundle of separate, divisible rights, each of which can be licensed, sold, or kept independently. Section 106 enumerates six exclusive rights, and a copyright owner can slice them as finely as a deli counter. Understanding the bundle is essential, because infringement is nothing more or less than the violation of one of these rights without permission and without a defense.

The first is the right to reproduce the work—to make copies (for sound recordings, "phonorecords"). This is the core right; photocopying a book, ripping a film, or duplicating a database all implicate it. The second is the right to prepare derivative works—new works based on the original, such as a film adaptation of a novel, a translation, a sequel, a remix, or an English-language version of a Spanish song. The third is the right to distribute copies to the public by sale, transfer, rental, lease, or lending. The fourth is the right to perform the work publicly—relevant to literary, musical, dramatic, and choreographic works, films, and the like. The fifth is the right to display the work publicly—relevant to literary, musical, dramatic, and graphic works, and to individual frames of a film. The sixth, added later and narrower, is the right to perform a sound recording publicly by means of a digital audio transmission—the right that underlies digital streaming royalties for recordings.

Each of these is bounded by limitations Congress wrote into Sections 107 through 122. The most famous limitation is fair use (Section 107), which gets its own section below. But others matter constantly. The first sale doctrine (Section 109) is the reason a bookstore can resell used books and you can give away a DVD you bought: once a particular lawful copy is sold, the copyright owner's distribution right in that copy is exhausted. (First sale governs the distribution of a physical copy, not the reproduction right and not, in the digital world, the making of new copies—which is why "reselling" an e-book or an MP3 is legally fraught.) There are compulsory licenses for making cover versions of musical compositions, exemptions for libraries and classrooms, and special rules for cable and satellite retransmission. The architecture is consistent: Section 106 grants a broad bundle, and the following sections carve out the public-policy exceptions.

Who Owns It: Authorship, Joint Works, and the Work-Made-for-Hire Trap

The default rule of ownership is simple and elegant: copyright vests initially in the author, meaning the human being (or beings) who actually created the work, the instant it is fixed. But three wrinkles complicate that default, and each is a frequent source of expensive disputes.

The first wrinkle is joint authorship. When two or more people create a work with the intent that their contributions merge into "inseparable or interdependent parts of a unitary whole," they are joint authors, and they co-own the entire work as tenants in common. The consequences surprise people. Each joint author may independently exploit the whole work or license it non-exclusively without the others' permission—subject only to a duty to account to the co-owners for their share of the profits. So if Alice and Ben write a song together as joint authors, Ben can license it to a commercial without asking Alice, though he must pay Alice her share. The intent requirement matters: courts (following influential cases like Childress v. Taylor and Aalmuhammed v. Lee) generally require that each putative joint author intended to be a co-author and contributed independently copyrightable expression—not just ideas or editorial suggestions. A researcher who hands a novelist facts is not a joint author; a co-writer who drafts whole scenes likely is.

The second wrinkle is the collective work, which is different from a joint work. In a collective work—a magazine, an anthology, an encyclopedia—a number of separate and independent contributions are assembled into a whole. The author of each contribution keeps copyright in their own piece, and the compiler owns a separate copyright in the selection, coordination, and arrangement of the whole (the Feist-style "thin" compilation copyright). Absent a transfer, the collective-work owner gets only the privilege of reproducing the contribution as part of that collective work and revisions of it. We treat this in detail in contributions to a collective work.

The third wrinkle is the big one: the work made for hire (often abbreviated WMFH). This is the most important exception to the "author owns it" rule, and it is a trap for the unwary on both sides of a creative transaction. When a work is "made for hire," the employer or commissioning party is deemed the author—not the human who created it—and owns the copyright from the start, for the full term, with no later termination right. The statute provides exactly two paths to work-made-for-hire status, and they are not interchangeable:

  1. Employee within the scope of employment. A work prepared by an employee within the scope of their job is automatically a work made for hire. The catch is the word "employee." In Community for Creative Non-Violence v. Reid, 490 U.S. 730 (1989), the Supreme Court held that "employee" is determined by common-law agency principles—a multi-factor test weighing who controls the manner and means of the work, who supplies the tools, the duration of the relationship, tax treatment, benefits, and more. A salaried staff designer is an employee; a freelancer you call a "contractor" almost certainly is not, no matter what your invoice says.

  2. Specially ordered or commissioned work, by written agreement, in one of nine categories. A commissioned work from a non-employee can be a work made for hire only if (a) the parties expressly agree in a signed writing that it is a work made for hire, and (b) the work falls into one of nine statutorily enumerated categories: a contribution to a collective work, part of a motion picture or other audiovisual work, a translation, a supplementary work, a compilation, an instructional text, a test, answer material for a test, or an atlas. Notice what is missing from that list: ordinary software (most code), a logo, a standalone photograph, a musical sound recording, a painting. If Acme Corp. hires a freelance developer to build its app and the contract says "this is a work made for hire," that magic phrase does nothing for the code, because software is not on the list of nine categories and the developer is not an employee.

This is the single most common copyright drafting mistake we see, and it is why every well-drafted creative-services contract includes a belt-and-suspenders assignment clause in addition to (or instead of) work-made-for-hire language: "and to the extent any deliverable does not qualify as a work made for hire, the contractor hereby assigns all right, title, and interest in it to the company." Get the WMFH question wrong and a company can spend years believing it owns its own software only to discover the freelancer does—a disaster we explore further in employee invention assignment agreements and in the IP layers of protectable content in mobile apps. One more crucial consequence: works made for hire have a different duration (below), and—because the company is the author—the powerful statutory termination right discussed next does not apply to them.

Duration: How Long Copyright Lasts

The single most-asked copyright question is also one of the trickiest, because the answer depends on when the work was created and, for older works, whether and how it was published. Here is the modern, post-1978 landscape, with the legacy rules kept brief.

For works created on or after January 1, 1978 by a known individual author, copyright lasts for the life of the author plus 70 years. For a joint work, the clock runs from the death of the last surviving author plus 70 years. For works made for hire and anonymous or pseudonymous works, the term is 95 years from first publication or 120 years from creation, whichever expires first. These terms reflect the Sonny Bono Copyright Term Extension Act of 1998, which added 20 years to existing terms and which the Supreme Court upheld against a constitutional challenge in Eldred v. Ashcroft, 537 U.S. 186 (2003). One welcome consequence of that 1998 extension expiring on a rolling basis is that, beginning in 2019, published works have been entering the public domain each January 1—1923's works in 2019, and so on—a yearly windfall for the commons (2024 famously freed the earliest Mickey Mouse from Steamboat Willie, subject to lingering trademark questions).

For pre-1978 works, the rules are a minefield best navigated with a chart (the Copyright Office and Cornell's public-domain guides maintain excellent ones). In broad strokes: works published in the United States before 1929 are in the public domain. Works published between 1929 and 1977 had a first term of 28 years and a renewal term (now extended to a total of 95 years from publication), but only if the required renewal was made—and an enormous number of mid-century works fell into the public domain because their owners forgot to renew. Renewal became automatic for works published 1964–1977, but for works published 1923–1963 a failure to renew was fatal. (We cover this in detail in renewal of copyright.) Works created before 1978 but never published before then generally follow the life-plus-70 rule, with statutory backstops ensuring such works did not expire before the end of 2002, and longer if published by then.

Two practical takeaways. First, copyright duration is long—routinely more than a century for corporate works—and it does not depend on registration or renewal for anything created since 1978. Second, public-domain status is a research question, not a guess; the cost of assuming a work is free when it is not can be statutory damages of up to $150,000 (below), so when in doubt, check the chart or ask counsel.

Transfers, Licenses, and the Termination Right of Section 203

Because the copyright bundle is divisible, an owner can transfer all of it or any slice of it. The Copyright Act draws a sharp line between two kinds of transfers. An exclusive license or assignment—any transfer of copyright ownership, including an exclusive license of even one of the six rights—is a "transfer of copyright ownership" and must be in writing and signed by the owner (or the owner's authorized agent) to be valid. A non-exclusive license, by contrast, can be granted orally or even implied from conduct: if you commission a photographer for your website and they deliver the photos knowing the use, courts may find an implied non-exclusive license even without a written word. Copyrights can also be inherited by will or pass under state intestacy law—a topic that intersects with estate planning in who will inherit your intellectual property.

Transfers can—and often should—be recorded with the Copyright Office. Recordation is not required for a transfer to be valid between the parties, but it confers real advantages: it provides constructive notice to the world (so later purchasers cannot claim ignorance) and it establishes priority among competing transfers, much like recording a deed in real estate. In a merger or acquisition, untangling who owns which copyrights and whether the chain of title was properly papered and recorded is a standard and sometimes harrowing diligence exercise.

Now for one of the most fascinating and counterintuitive provisions in all of intellectual property: the termination of transfers right under Section 203. Congress worried that authors—often young, unknown, and without bargaining power—would sell their copyrights for a pittance before anyone knew the work would become valuable. So it gave authors and their heirs an inalienable right to terminate a transfer or license and reclaim the copyright, regardless of what the contract says. The right is unwaivable; an author cannot sign it away in advance, no matter how the lawyers paper it. For transfers made on or after January 1, 1978, termination may generally be effected during a five-year window beginning 35 years after the grant (or, for grants covering publication rights, 35 years after publication or 40 years after the grant, whichever is earlier). The author or heirs must serve an advance written notice of termination on the grantee within prescribed time windows (typically two to ten years before the chosen effective date) and record a copy with the Copyright Office.

The consequences are dramatic and very real in 2026. The recording artists and songwriters whose careers began in the late 1970s and 1980s—and the heirs of those who have since died—have been filing termination notices to reclaim master recordings and song catalogs from the labels and publishers who bought them decades ago, producing a steady stream of high-stakes litigation over notice timing, what counts as a "grant," and the perennial WMFH escape hatch (remember: works made for hire have no termination right, so labels routinely argue that recordings were works made for hire, while artists argue they were not). The termination right is a deliberate, structural thumb on the scale in favor of creators, and it is one of the reasons a copyright assignment is never quite as permanent as it looks. (Note that older transfers of pre-1978 copyrights are governed by a parallel provision, Section 304, with its own timing.)

Fair Use: The Great Safety Valve After Google and Warhol

No copyright concept is invoked more often or understood less well than fair use. Fair use is the doctrine, codified in Section 107, that permits certain unauthorized uses of copyrighted works without liability, in service of the First Amendment values and the "promote the Progress" purpose baked into the Constitution. It is a defense to infringement, and—critically—it is decided case by case, by weighing four non-exclusive statutory factors. There is no bright line, no "if you use less than 10% it's fine," no "non-commercial means fair," no "giving credit cures it." Those are myths. Fair use is a judgment call, and reasonable judges disagree.

The four factors are:

  1. The purpose and character of the use, including whether it is commercial and—above all in modern doctrine—whether it is transformative, meaning it adds something new, with a further purpose or different character, altering the original with new expression, meaning, or message. Transformativeness, a concept the Supreme Court embraced in Campbell v. Acuff-Rose Music, Inc., 510 U.S. 569 (1994) (the 2 Live Crew "Oh, Pretty Woman" parody case), became the gravitational center of fair use for two decades.

  2. The nature of the copyrighted work—fact-based and published works are more susceptible to fair use than highly creative or unpublished ones.

  3. The amount and substantiality of the portion used—both quantitatively and qualitatively; copying a small but "heart of the work" portion can weigh against fair use.

  4. The effect on the potential market for or value of the original—does the use substitute for the original or usurp a market (including licensing markets) the owner could exploit? This factor is often called the most important.

Two recent Supreme Court decisions have reshaped how these factors interact, and any current understanding of fair use must reckon with both.

In Google LLC v. Oracle America, Inc., 593 U.S. 1 (2021), Google had copied roughly 11,500 lines of Oracle's Java API "declaring code"—the labels and organization that let programmers call pre-written functions—to build the Android platform, so that Java developers could use familiar commands. After more than a decade of litigation, the Supreme Court held Google's copying was fair use as a matter of law. The Court assumed (without deciding) that the declaring code was copyrightable, then ran the four factors. It treated the nature of the work as weighing for fair use because declaring code is "inherently bound together with uncode­able ideas" and functional. It found the purpose transformative because Google used the interface to create a new platform for a new computing environment (smartphones). And on market harm, it credited findings that Java was not well positioned for the smartphone market and that enforcing Oracle's claim would risk locking up the building blocks of new creation. Google v. Oracle is, at bottom, a fair-use decision deeply concerned with the functional, interoperable nature of software—a modern echo of Baker v. Selden's instinct that copyright should not become a patent on methods of operation. It is essential reading for anyone working with software interfaces; see also legal protection of software.

Then, in Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, 598 U.S. 508 (2023), the Court reined in the runaway expansion of "transformativeness." Photographer Lynn Goldsmith had taken a portrait of the musician Prince. Andy Warhol used that photograph to create a series of silkscreen images; years later, after Prince's death, the Warhol Foundation licensed one of those images ("Orange Prince") to Condé Nast for a magazine cover. Goldsmith, who licensed her own photographs to magazines, sued. The Foundation argued the Warhol image was transformative—it conveyed a different meaning, commenting on celebrity. The Supreme Court, in an opinion by Justice Sotomayor, disagreed on the first factor as applied to this specific commercial licensing use. The Court held that the first factor asks not merely whether a new work has a different meaning or message, but whether and to what degree the use has a different purpose or character, and that this must be weighed against the commercial nature of the use and the degree of transformation. Crucially, because both Goldsmith's photo and the licensed Warhol image served the same essential purpose—illustrating a magazine story about Prince—and the use was commercial, the first factor favored Goldsmith. A new meaning or message, the Court warned, is not enough on its own; otherwise every derivative work (which by definition recasts the original) would escape the copyright owner's exclusive right to prepare derivatives.

The lesson practitioners draw from Warhol is that fair use after 2023 places renewed emphasis on whether the challenged use competes with or substitutes for the original in the same market, and that a claim of "I transformed it" will not carry the day if your use serves the same commercial function the original serves. Together, Google and Warhol bracket modern fair use: software interoperability and the creation of genuinely new platforms lean toward fair use; commercial reuse of expressive art that competes in the original's licensing market leans against it. These principles are at the white-hot center of current litigation over whether training generative AI models on copyrighted works is fair use—a fight playing out in cases brought by authors, artists, and the news media, which we track in copyright infringement claims against generative AI.

A practical warning to close this section: because fair use is a fact-intensive defense decided only after litigation begins, "it's probably fair use" is a risk position, not a permission slip. When the stakes are high, the right move is often to license, to seek counsel, or to design the use to maximize transformation and minimize market substitution—not to gamble on four factors a court will weigh after the fact.

Registration and the Fourth Estate Rule

Here is the paradox at the heart of American copyright. Copyright protection is automatic—you own it the moment you fix the work, with no registration required. And yet registration is enormously important, to the point that for U.S. works it is a practical prerequisite to enforcing your rights at all. Both statements are true, and the reconciliation is the key to copyright strategy.

Registration is a formal claim filed with the Copyright Office. It requires three things: a completed application, a filing fee, and a deposit of the work (the "deposit copy" the Office keeps). You can file online through the eCO system at copyright.gov, which is faster and cheaper than paper. The effective date of registration is the day the Office receives all three required elements in acceptable form—not the day it finishes examining your claim, which can take many months. Our comprehensive registration guide walks through the mechanics, and the step-by-step how-to covers the application screens; specialized companions address registering computer programs, photographs, websites and website content, books and manuscripts, games, and derivative works.

Why bother, if rights are automatic? Because the statute loads registration with benefits, three of which are decisive:

First, registration is a precondition to suit for U.S. works. Section 411(a) requires that a U.S. work be registered (or registration refused) before the owner can file an infringement lawsuit. For years, courts split over whether "registration" meant applying or actually obtaining the Office's decision. The Supreme Court resolved it unanimously in Fourth Estate Public Benefit Corp. v. Wall-Street.com, LLC, 589 U.S. 1 (2019), holding that "registration . . . has been made" means the Copyright Office has acted on the application—approving or refusing it—not merely that the owner has filed it. In plain terms: you generally cannot sue until the Copyright Office has ruled on your application. Because ordinary processing can take many months, Fourth Estate makes early registration a litigation necessity. A creator who waits until infringement strikes faces a painful choice: wait out the queue, or pay for special handling (expedited processing for an extra fee) to register in days rather than months. The smart practice—register valuable works promptly, before any dispute—exists largely because of this rule.

Second, timely registration unlocks statutory damages and attorney's fees. Under Section 412, an owner can recover statutory damages (a court-set amount per work infringed, with no need to prove actual loss) and attorney's fees only if the work was registered before the infringement began, or, for a published work, within three months of first publication. Miss that window and you are limited to actual damages and the infringer's profits—which can be hard to prove and may be modest, while your legal fees mount with no prospect of recovery. This single provision is why intellectual-property lawyers preach early registration like a gospel: it transforms the economics of enforcement.

Third, registration is evidence. A certificate of registration made before or within five years of first publication constitutes prima facie evidence of the validity of the copyright and of the facts stated in the certificate, shifting the burden to the infringer to prove otherwise. Registration also lets an owner record the copyright with U.S. Customs and Border Protection to block importation of infringing copies—useful against counterfeit goods.

Two related concepts round out the registration picture. A copyright notice—the © symbol (or "(p)" for sound recordings), the year of first publication, and the owner's name—has not been required since the United States joined the Berne Convention, effective March 1, 1989. But it remains worthwhile: a proper notice forecloses an "innocent infringement" defense that could otherwise reduce damages, identifies the owner for those seeking permission, and helps prevent a work from becoming an "orphan." We cover the details in copyright notice: form, function, and best practices. And mandatory deposit (Section 407) requires that two copies of the "best edition" of most works published in the United States be deposited with the Library of Congress within three months of publication—an obligation distinct from registration deposit, enforceable by fines, though in practice satisfied for most registrants when they register.

Infringement and Remedies

To win a copyright infringement suit, a plaintiff must prove two things: (1) ownership of a valid copyright, and (2) copying of original elements of the work. The first element is usually established by the registration certificate. The second is where the action is, because direct evidence of copying (a confession, a smoking-gun email) is rare. Plaintiffs typically prove copying circumstantially by showing the defendant had access to the work plus substantial similarity between the works—similarity not just in unprotectable ideas or facts but in the protected expression. Courts use various tests to filter out the unprotectable elements (the Feist and Baker leftovers—ideas, facts, scènes à faire or stock elements, merged expression) before comparing what remains. If two songs share only a common chord progression, or two software programs share only the functions dictated by the task, there is no infringement, however similar they feel. The independent-creation defense is absolute: if the defendant created the same thing without copying, there is no liability, because copyright (unlike patent) protects only against copying, not against parallel creation.

Liability extends beyond the direct copier. Contributory infringement reaches one who knowingly induces or materially contributes to another's infringement, and vicarious infringement reaches one who has the right and ability to supervise the infringing activity and a direct financial interest in it. These secondary-liability doctrines are what put platforms, app stores, and service providers in the crosshairs—and what makes the DMCA safe harbors, discussed next, so valuable to them.

The remedies arsenal is formidable. A successful plaintiff may obtain injunctions (court orders to stop infringing and to impound or destroy infringing copies); actual damages plus the infringer's profits attributable to the infringement (with the owner needing only to prove gross revenue, after which the infringer bears the burden of proving deductible costs and the portion of profit attributable to other factors); or, if the registration timing of Section 412 was satisfied, statutory damages. Statutory damages run from $750 to $30,000 per work infringed at the court's discretion, rising to as much as $150,000 per work for willful infringement and dropping to as little as $200 for innocent infringement. The per-work multiplier is what makes statutory damages so fearsome: an infringer who copied fifty registered photographs faces a theoretical exposure into the millions before a single dollar of actual harm is proven. Prevailing parties may also recover attorney's fees under Section 505 (again gated by Section 412 timing), and the Supreme Court has confirmed that the recovery period for damages can reach back more than three years for timely-filed claims under the discovery rule (Warner Chappell Music, Inc. v. Nealy, 601 U.S. 366 (2024)). Criminal penalties exist for willful infringement on a commercial scale. For the practical and human cost of infringement—and a tour of the penalties—see what are the consequences of pirating intellectual property, and for the dollars-and-cents picture across IP, damage statistics for intellectual property litigation.

One procedural innovation deserves mention: the Copyright Claims Board (CCB), a small-claims tribunal within the Copyright Office created by the CASE Act of 2020 and operational since 2022. It offers a voluntary, low-cost forum for disputes with damages capped at $30,000, designed for individual creators—photographers, illustrators, small businesses—who cannot justify the expense of federal litigation. Respondents may opt out, which limits its reach, but it has quietly become a real venue for the long tail of copyright disputes.

The DMCA: Safe Harbors and Anti-Circumvention

The Digital Millennium Copyright Act of 1998 (DMCA) bolted two important structures onto the Copyright Act for the internet age, and they pull in opposite directions—one shields intermediaries, the other arms rightsholders.

The first is the set of safe harbors in Section 512. The internet could not function if every platform were strictly liable for every infringing file a user uploaded. The DMCA safe harbors give qualifying online service providers immunity from monetary liability for user-generated infringement, provided they meet conditions: they must designate a registered agent to receive complaints, adopt and reasonably implement a repeat-infringer policy, lack actual knowledge or "red flag" awareness of specific infringement, not receive a financial benefit directly attributable to infringement they can control, and—above all—expeditiously remove or disable access to allegedly infringing material upon receiving a proper takedown notice. This is the famous notice-and-takedown system: a rightsholder sends a Section 512(c)(3) notice, the platform takes the material down, and the user may file a counter-notice to have it restored if they believe the takedown was mistaken. The system is not without abuse; Section 512(f) creates liability for knowing material misrepresentations in a takedown, and in Lenz v. Universal Music Corp. (the "dancing baby" case) the Ninth Circuit held that a sender must form a good-faith belief that the use is not fair use before sending a notice. We cover the mechanics in how to file a DMCA takedown notice and respond to one and the provider-side requirements in depth in DMCA safe harbors for online service providers. Note that the DMCA safe harbors address copyright specifically and are distinct from the broader (and more contested) immunity of Section 230 of the Communications Decency Act, which we examine in Section 230 reform and platform liability.

The second structure is the anti-circumvention regime of Section 1201, which makes it unlawful to circumvent technological protection measures ("digital locks") that control access to copyrighted works, and to traffic in tools designed to do so. This is the law behind region-coded DVDs, copy-protected software, and the recurring fights over whether you may bypass the locks on a device you own. Because Section 1201 can sweep in legitimate activity (security research, repair, accessibility), the law provides for triennial rulemakings in which the Copyright Office grants temporary exemptions—the mechanism that, for example, has at times allowed consumers to unlock their phones, as we discuss in unlocking cell phones: the law and the DMCA exemption. Section 1201 sits at the contested intersection of copyright and the right-to-repair movement and is, in practice, less about copying than about control of access.

The International Picture: Berne and the Myth of "International Copyright"

A persistent misconception deserves a clean burial: there is no such thing as an international copyright that protects a work everywhere at once. Copyright is territorial. Protection in any given country depends on that country's national law. What makes global protection workable is a web of treaties under which countries agree to protect one another's authors.

The cornerstone is the Berne Convention for the Protection of Literary and Artistic Works (1886), which the United States finally joined effective March 1, 1989—a date you will recognize because it is when U.S. copyright notice became optional. Berne rests on two principles. The first is national treatment: each member country must give foreign authors from other member countries the same protection it gives its own nationals. So a French novelist's book is protected in the United States on the same terms as an American's, and vice versa. The second is a prohibition on formalities: Berne forbids member countries from conditioning the existence or exercise of copyright on formalities like registration or notice for works originating in other member countries. That is why the United States had to make notice optional and why—a quirk worth knowing—the Section 411(a) registration requirement to sue applies only to U.S. works: foreign Berne works can be enforced in U.S. courts without first registering, because requiring registration would violate Berne. The United States is also party to the TRIPS Agreement (administered through the World Trade Organization, adding enforcement obligations and dispute resolution) and the WIPO Copyright Treaty (which addressed digital-era rights and underlies parts of the DMCA). The practical upshot for a creator: your work enjoys protection in most of the world automatically by virtue of these treaties, but the scope, term, and remedies are determined country by country, so cross-border enforcement is a multinational exercise, not a single global filing.

A Worked Example: Following One Work Through the System

Abstractions become clear when a single work runs the gauntlet. Meet Bluebird Records, a small independent label, and Mira Okafor, an emerging songwriter and recording artist.

Mira writes a song—words and music—in her apartment, recording a rough demo on her phone. The instant that demo saves, two copyrights spring into existence: one in the musical composition (the underlying notes and lyrics, a "musical work") and one in the sound recording (the fixed performance). Both are original (independently created, with creativity well above the Feist floor) and fixed. She owns both, automatically, with no paperwork. The chord progression she used, however—a common one—belongs to no one; that is an unprotectable building block under Baker and merger principles, so another songwriter using the same progression infringes nothing.

Mira signs with Bluebird. The contract assigns the sound-recording copyright to the label (in a signed writing, as Section 204 requires) and grants Bluebird an exclusive license to administer the composition, while Mira keeps her songwriter's share. The label, being prudent, records the assignment with the Copyright Office to establish priority and registers the recording promptly—well within three months of release—so that if a piracy network rips the track, Bluebird can pursue statutory damages and attorney's fees under Section 412 and can file suit immediately once the Office acts, satisfying Fourth Estate.

A year later, a popular YouTuber uses thirty seconds of Mira's recording as background to a cooking video. Is it fair use? Run the factors. The use is commercial and not obviously transformative—it serves an entertainment/illustrative purpose much like the original's, the Warhol red flag. The recording is creative (factor two against fair use). Thirty seconds may be quantitatively small but could capture the "heart" of the song. And it arguably substitutes for a sync license Bluebird would otherwise sell. On balance, this leans against fair use, so Bluebird's options are a license or a takedown. It sends a DMCA Section 512 notice; YouTube, protected by the safe harbor so long as it acts, removes the clip; the YouTuber, believing the use fair, files a counter-notice—and now the parties either negotiate a license or litigate the fair-use question.

Thirty-five years on, suppose the composition has become a standard worth millions. Mira (or her heirs) can serve a Section 203 notice of termination to claw back the rights she granted Bluebird back in her unknown days—an inalienable right Bluebird cannot have bargained away, and one that does not apply to the sound recording if a court were to find it a work made for hire (which is exactly what the parties would fight about). One small song, and it has touched fixation, originality, the idea-expression line, the divisible bundle, assignment formalities, recordation, registration timing, statutory damages, fair use after Warhol, the DMCA, and the termination right. That is the copyright system in miniature.

Copyright in Its Family: How It Differs From Trademark and Patent

People conflate the three big federal IP rights constantly, often calling everything a "copyright." They protect different things and arise in different ways. Copyright protects original expression—the way an idea is written, drawn, coded, performed, or composed—arises automatically on fixation, and lasts roughly a lifetime plus 70 years. Trademark protects source identifiers—brand names, logos, slogans, and trade dress that tell consumers who stands behind a product—arises from use in commerce (strengthened by registration), and can last indefinitely so long as the mark is used and policed. Patent protects functional inventions and ornamental designs, must be applied for and examined, and lasts a limited term (twenty years for utility patents from filing). A single product routinely carries all three: a sneaker might have a utility patent on its sole technology, a design patent and trade dress on its shape, a trademark on the swoosh, and a copyright on the artwork printed on the box and the code in its companion app. For the full comparison see copyright vs. trademark vs. patent and the two-way primer copyright vs. trademark: what is the difference. The cardinal rule: pick the right tool for the right asset, and layer them where the law allows.

Key Takeaways

Copyright is automatic, but enforcement is not. Protection attaches the instant an original work is fixed, yet for U.S. works you generally cannot sue until the Copyright Office has acted on a registration (Fourth Estate), and you forfeit statutory damages and attorney's fees unless you registered before the infringement or within three months of publication (Section 412). The single best thing most creators and businesses can do is register valuable works promptly.

Copyright protects expression, never ideas, facts, systems, or methods (Baker v. Selden, Section 102(b)), and demands a genuine spark of originality, not mere effort (Feist). It confers a divisible bundle of six exclusive rights (Section 106), each licensable separately. Ownership defaults to the human author, but the work-made-for-hire rules and the formalities of written transfers can move it—so paper your assignments carefully, and remember the inalienable Section 203 termination right that lets authors reclaim rights after 35 years. Fair use is a fact-bound, four-factor defense, narrowed by Warhol where a use competes commercially with the original and expanded by Google v. Oracle for functional, interoperable software. The DMCA shields compliant platforms through notice-and-takedown while policing circumvention of digital locks. And copyright is territorial, made global only through treaties like Berne. Master these moving parts and the rest of copyright—from registration mechanics to the DMCA—falls into place.

Frequently Asked Questions

Do I have to register my work to have a copyright? No. Copyright attaches automatically the moment your original work is fixed in a tangible medium. But registration is essential in practice: for U.S. works it is a precondition to filing an infringement suit (the Fourth Estate rule), and registering before infringement (or within three months of publication) is the only way to unlock statutory damages and attorney's fees. Register early; do not wait for trouble.

Can I copyright my idea, my business name, or my recipe? Not as a copyright. Copyright protects expression, not ideas, systems, methods, or facts (Section 102(b); Baker v. Selden). A business name or slogan is too short for copyright but may be protectable as a trademark. A recipe's bare list of ingredients and functional steps is an unprotectable method, though original creative writing surrounding it can be copyrighted. The underlying idea is always free for others to express in their own way.

Is it fair use if I only use a little, give credit, or don't make money? Those are common myths. Fair use is decided by weighing four factors—purpose and character (including transformativeness), nature of the work, amount used, and market effect—case by case. There is no percentage that is automatically safe, crediting the author does not cure infringement, and even non-commercial use can fail if it substitutes for the original. After Warhol v. Goldsmith, a use that competes commercially in the original's market is especially vulnerable.

How long does copyright last? For works created on or after January 1, 1978, the term is the life of the author plus 70 years; for works made for hire and anonymous/pseudonymous works, 95 years from publication or 120 years from creation, whichever is shorter. Pre-1978 works follow older rules that can turn on publication and renewal—check a duration chart or ask counsel before assuming a work is in the public domain.

Who owns a copyright in work I paid a freelancer to create? Not necessarily you. Absent a valid transfer, the creator owns it. Calling it a "work made for hire" in a contract only works if the creator is your employee or the work falls into one of nine narrow statutory categories—ordinary software, logos, and standalone photos are not on that list. Always include a written assignment clause so ownership passes to you regardless. See our article on employee invention assignment agreements.

Does AI-generated content have a copyright? Purely machine-generated output with no meaningful human authorship is not registrable, per the Copyright Office and Thaler v. Perlmutter. A human must contribute original creative expression for a work made with generative AI tools to qualify, and an applicant must disclaim the AI-generated portions. This area is evolving quickly; see our overview of AI legal issues.

Is my U.S. copyright good in other countries? There is no single "international copyright," but thanks to treaties—chiefly the Berne Convention, plus TRIPS and the WIPO Copyright Treaty—your work is generally protected in most countries on the same terms those countries give their own authors (national treatment). Scope, term, and remedies vary by country, so enforcement abroad is country-by-country.

What's the difference between the DMCA and Section 230? The DMCA's Section 512 safe harbors specifically govern copyright infringement by users, conditioned on notice-and-takedown and a repeat-infringer policy. Section 230 of the Communications Decency Act is a broader immunity for platforms regarding most user content, but it expressly excludes intellectual property claims—which is precisely why the DMCA exists to fill the copyright gap. See Section 230 reform and platform liability.

Related Articles

This article is general information, not legal advice. Copyright law is fact-specific and evolving; for guidance on your particular situation, consult qualified counsel.