Copyright has become the battleground à la mode for AI-related litigation. Referring to the technology as “nothing more than a plagiarism machine” developed by “greedy and craven companies who want to take human talent out of entertainment,”1 artists and creators have increasingly called for copyright-driven crackdowns on the generative AI industry.2 In the courts, parties have filed lawsuits against major AI companies like OpenAI,3 Stability AI,4 and Meta,5 and stakeholders from all sides have appealed to the U.S. Copyright Office6 in response to its “study of the copyright law and policy issues raised by [AI].”7 Among those responses were comments “lament[ing] an imagined future without new human-authored works”8 and “voic[ing] concerns that generative AI will prevent artists’ fair renumeration [sic] for their work.”9
Copyright was “designed to protect originality [and] creativity,”10 so perhaps it’s “natural” to see copyright law as a golden hammer for protecting artists.11 But commentators have grown weary of copyright as a cure-all. Copyright litigation has ultimately produced little guidance, at least in the way of judicial holdings, and fundamental questions — like whether AI-generated works should be copyrightable — are still subject to debate.12 For some, this lack of clarity suggests that concerns about generative AI ought to be addressed with other legal and policy tools that can better respond to the technology. Scholars have suggested, for example, “that trying to fit the tension created by [generative AI] into the copyright framework can put at risk already strained concepts like fair use.”13 The Electronic Frontier Foundation (EFF) likewise has remarked that while copyright is “supposed to encourage new creativity” when it’s “[d]one right,” “[s]tretching it” to deploy against generative AI “would have the opposite effect.”14
That language — “stretching” — is useful for understanding copyright’s weaknesses. Much of the doubt about the future of generative AI and copyright comes from uncertainty about whether copyright’s fundamental principles, like originality and creativity, can (or should) accommodate the technology.15 That is, can AI create original works? Can it be creative?
These questions are difficult for myriad reasons, but they’re made more complicated by the fact that copyright’s originality and creativity requirements have — over time — largely been left to waste away. The Supreme Court has frequently affirmed that the requirements pose only a low bar to copyrightability,16 and “[s]eeing the depths to which the creativity requirement has sunk, some call for abandoning it altogether.”17 But “understanding what creativity is . . . becomes a key consideration in debates about” generative AI.18 And some have argued that “this is [a] dimension where copyright principles might not be reconcilable with the challenges posed by [generative AI] without having to twist their long-standing definition and interpretation in a forceful way.”19 In other words, without a clear understanding of what “originality” and “creativity” really mean, it will be all but impossible to use them to understand what generative AI means for copyright — and for parties concerned about generative AI to use copyright in defense.
But this Note suggests that copyright — and originality, in particular — can still be useful to those seeking to leverage it against generative AI and, in fact, that it’s a simple fix. Although the requirements are minimal, originality and creativity have remained fundamental to copyright; they need only to be revitalized. Even when organizations like the EFF acknowledge the “Exciting Things About AI Art Generation,” their optimism often includes a caveat: that AI “leads to a different kind of ‘creativity’ than human artists produce on their own”20 — and perhaps a different kind than copyright was meant to protect. Making the language of originality useful requires, at least as a first step, nothing more than giving voice to what “kind of creativity” that is. And fortunately, the answer can be found in long-held theories of property and personhood and in the history of originality itself. In this sense, this Note differs from scholarship urging that the originality requirement be made more complicated or more stringent.21 Indeed, it seeks not to comment on whether the bar to originality is too high or too low but only to suggest that we know one thing for certain about where that bar should be: smack-dab between human- and AI-generated works.
While today’s originality requirement — satisfied by only a “modicum of creativity”22 — offers an inadequate shield against threats posed by generative AI, its now-neglected history offers some reinforcement. Though the bar for originality has always been low, a shift in vocabulary — initiated by Justice Holmes’s opinion in Bleistein v. Donaldson Lithographing Co.23 — untethered copyrightability from its roots in human creativity and shifted it toward its modern emphasis on commercial value, erasing consideration of “personality.” Today’s copyright law simply takes for granted that aesthetic creation is — and has historically been understood as — a human endeavor. Now, without the contours of personality to give it shape, originality in copyright is, as a matter of principle, perfectly capable of accommodating AI creativity: Attempts to exclude AI-generated works in this framework will likely remain an unsatisfying exercise in line-drawing. A return to the pre-Bleistein, personality-driven conception of originality, by contrast, would do two things. First, and most elementally, it would revitalize the originality requirement, beefing it up to make it more useful in the era of generative AI. In so doing, it would more clearly distinguish originality from authorship and help explain some of the tensions that color today’s debates about the technology. Second, it would offer litigants concerned about AI the language to talk about harms in a way that better reflects what may otherwise be disregarded as a legally irrelevant sentiment: that we just don’t want (and, in fact, never have wanted) to protect works without personality — works that aren’t human. It better explicates what is, at base, an emotional connection to human creativity.
This Note tells the story of personality in copyright and shows what recovering personality can add to conversations about the changing relationship between copyright and generative AI. It proceeds in four parts. Part I provides a brief introduction to copyright’s originality requirement. Part II addresses the “human requirement” in copyright through the lens of Hegelian property theory and recounts how, in one fell swoop — that is, in one brief opinion — personality was both brought to light and quickly overshadowed as a “juridically significant concept.”24 Part III engages in a sort of revisionist history, showing how personality nevertheless appears in and elucidates existing copyright law. Finally, Part IV clarifies how a personality-driven originality inquiry differs in purpose from the authorship analysis, and illustrates how personality can be used by litigants in today’s generative AI cases to advocate for compensation for harms felt not only on economic axes but also on the fundamental level of personhood.
At bottom, this Note’s argument is a modest one: that originality requires personality and, thus, that it assumes personhood. But recovering the language of personality is powerful because it gives that language back to people who are trying to explain why copyright matters to them in the age of generative AI. In doing so, it also strengthens arguments in favor of more robust moral rights and economic guarantees for human creators in the AI era.
I. What Is Originality?
Originality occupies an unusual, if not paradoxical, position within copyright. It is “the bedrock principle” and “the very premise of copyright law,”25 but it poses only a relatively small speedbump to registrants on the road to copyrightability. Indeed, originality requires only “independent creation plus a modicum of creativity,”26 and the Supreme Court has made clear that “the requisite level of creativity is extremely low.”27 “The vast majority of works,” the Court has explained, will “make the grade quite easily.”28 And when originality does pose an obstacle, it’s often because a work fails to cross the line from mere facts or ideas to the level of expression required by copyright.29 Whether an expressive work is sufficiently “creative” as a matter of aesthetic merit is, in other words, rarely the load-bearing question of originality30 — and beyond determining that a work “possesses at least some minimal degree of creativity,”31 judges will explicitly avoid weighing in on the creative value of a work in the name of aesthetic neutrality.32
Perhaps because of its perception as a generally toothless standard,33 originality has been largely neglected in today’s most prominent debates about the scope of copyright protection for AI-generated works, which have focused instead on the boundaries of authorship. When “originality” is brought into the fold, it’s frequently baked into the authorship inquiry rather than addressed as a distinct requirement. Last year, for example, the Copyright Office affirmed its refusal to register Ankit Sahni’s AI-generated artwork, for which he had identified both himself and the “RAGHAV Artificial Intelligence Painting App” as coauthors.34 The Office began its decision with a restatement of the Copyright Act,35 explaining that it could register only “original works of authorship fixed in any tangible medium of expression.”36 The discussion that followed, however, addressed only one of the requirements enveloped in that phrase: It focused entirely on the authorship question, asking not whether the work was “original” or “fixed” at all.37 But though the Copyright Office never asked the originality question of Mr. Sahni’s work, it still affirmed that it “d[id] not contain enough original human authorship to support a registration.”38 Originality, in other words, seems to have been taken for granted in the Office’s authorship analysis: A deficiency in authorship engendered a deficiency in originality, too.
The problem with this kind of collapsed analysis is that it wilts originality. Authorship may “presuppose a degree of originality,”39 but letting originality fade into the peripheries of the authorship inquiry suggests that — as an independent requirement — it is ineffectual, at best, and entirely duplicative, at worst. But can originality — the core of copyright — really have diminished into a doctrinal red herring? One should hope not. While the originality requirement “derives from the fact that, constitutionally, copyright protection may be claimed only by ‘authors,’”40 it can and should pull its own weight. And indeed, retelling the history of originality shows that it can. While both standards prioritize human-generated works, originality better reflects a cultural commitment to human creativity. Copyright’s originality requirement embodies an independent set of values — and can perhaps achieve a different purpose — than the precepts of authorship.
II. Copyrightability and “The Human Requirement”
That copyrightability depends on humanness is well-established.41 Indeed — though the so-called “human authorship requirement” is not an explicit constitutional mandate42 — the issue of humanity has long been fundamental to copyright’s authorship inquiry.43 The Copyright Act protects “original works of authorship,”44 and the Copyright Office has determined definitively that “[t]o qualify as a work of ‘authorship,’ a work must be created by a human being.”45 The Compendium III of U.S. Copyright Office Practices elaborates that “copyright law only protects ‘the fruits of intellectual labor’ that ‘are founded in the creative powers of the mind,’”46 and “[b]ecause copyright law is limited to ‘original intellectual conceptions of the author,’ the Office will refuse to register a claim if it determines that a human being did not create the work.”47 Courts have reinforced this standard for decades, refusing copyright protection for works purportedly created by non-humans, including spiritual beings,48 animals,49 and forces of nature.50 For any work to be copyrightable, registrants must show “some element of human creativity.”51 Today, the copyrightability of AI-generated works likewise hinges on the level of human involvement.52
But what if “the human question” were asked not only of the authorship requirement but of originality, too? While the human requirement has come to define copyright’s authorship analysis, it has been largely ignored in the law’s accompanying inquiry into originality. But that hasn’t always been the case. This section considers existing theories that account for the relationship between property and personhood and recounts the rise and fall of personality — a distinctly human characteristic — in Justice Holmes’s Bleistein opinion, explaining how the principle came to be so immediately overshadowed by the notion of commercial progress. In doing so, it demonstrates that originality, too, has always had a “human requirement.” Recovering personality today, in other words, is not about adding anything new to copyright; it is, instead, only about recognizing what has long been taken for granted.
Originality was once framed not only as a matter of creativity but of personality — a characteristic understood as inherently and uniquely human. Thus, while copyright today speaks less in the language of personality, its history suggests that “the human question” is not so much absent from the originality requirement as it is assumed. For decades, this hasn’t been a problem; the standards of authorship alone have offered a sufficient check on copyrightability. But now, with generative AI pushing the boundaries of creatorship, recovering personality — and the very human qualities it represents — may offer the reinforcement that copyright needs: a principled and historically-grounded distinction between human- and AI-generated works that can carry the emotional, cultural, and philosophical weight of the role that arts play in society.
That humanity — or what Professor Margaret Radin refers to as “personhood”53 — is fundamental to understanding property rights is not a novel proposition, nor one confined to the copyright context. Indeed, in her seminal article, Property and Personhood, Radin argues for a “personhood perspective” on property generally, based on the notion of personhood that Hegel developed in his Philosophy of Right.54 For Hegel, and for Radin, “the person becomes a real self only by engaging in a property relationship with something external.”55 Property is, for these thinkers, an embodiment of the owner’s will; it is an opportunity for a person to “take free will from the abstract realm to the actual”56 and thus to become a “fully developed individual.”57 When others acknowledge one’s property rights, it is a recognition not only of their possession of the thing but also of their “occupancy” of it,58 and in that sense, the recognition of property rights is the “acknowledgment that an individual is recognized by others as a person”59 — the acknowledgment of the worth of their will. In this sense, although Lockean property theory is often invoked as the basis of American property rights, Hegel’s view offers a more robust foundation for understanding the role of personal expression in property than a pure labor theory can provide.60
In copyright, this intuition has been given voice — at least briefly — in the language of “personality.”61 Personality’s role in the story of American copyright came to a head in Justice Holmes’s 1903 opinion in Bleistein v. Donaldson Lithographic Co. In that case, the Court was asked whether three “picture-posters” used as circus advertisements62 were copyrightable, or whether they were unfit for such protection for lacking sufficient “artistic merit or value.”63 Ultimately, the Court concluded that the posters were “entitled to the protection of the law” because “their worth . . . [was] sufficiently shown by the desire to reproduce them without regard to the plaintiffs’ rights.”64 This conflation of aesthetic and commercial value as an indication of copyrightability — and the ensuing commitment to judicial aesthetic neutrality65 — is the proposition for which Bleistein is often cited. But earlier in his brief opinion, Justice Holmes described copyrightable works another way: as “the personal reaction[s] of an individual upon nature.”66 He explained that “[p]ersonality always contains something unique” and that even “a very modest grade of art has in it something irreducible, which is one man’s alone.”67 For Justice Holmes, this “something irreducible” — personality — is what made a work original.68
Defining originality this way — as a product of personality — reflects an intuition about intellectual property, generally, and copyright, in particular: that the subject of these rights is somehow connected to the one who created it, that it’s a part of that person. This is the same intuition embodied in Radin’s “intuitive view of property for personhood.”69 As Professor Justin Hughes, expanding on Radin’s work, puts it, intellectual property differs even from other forms of property that carry sentimental value, like a wedding ring or a family home.70 While these objects can become “bound up with personhood,”71 they do so only because “an individual’s personality seems to move into an existing object.”72 Property rights in these things may reflect a personhood interest, but it is one based on ownership rather than creation — any deposit of “personality” happens post hoc, once the object has already come to be. By contrast, “[w]hen we first encounter a res of intellectual property . . . we may note that an individual’s personality caused the object to come into existence. The object comes into the world already an embodiment or reflection of some particular individual.”73 If the law is meant to reflect this intuition, then personality cannot be merely a characteristic of copyright, somehow identified in works after the fact. Personality must be, by contrast, the sine qua non of originality. It must exist independent of and prior to any work in which it may later be imbued. It must belong not to a work at all but to a human.
When Justice Holmes invoked personality as the requisite of originality, he likely intended to voice this very intuition. While today, we tend to “link, if not blur, our notions of creativity, originality, and personal expression,”74 Justice Holmes was operating squarely “within the tradition of nineteenth-century American literary romanticism.”75 Indeed, his “restatement of the originality requirement in terms of ‘personality’” was not drawn from thin air; rather, it “made explicit themes [already] latent in nineteenth-century case law and commentary.”76 As Professor Barton Beebe explains, “[w]ell-informed readers of the Bleistein opinion in 1903 would have . . . recognized” the “influence” of contemporary thinkers like Ralph Waldo Emerson, Walt Whitman, and John Dewey, who themselves reflected the “American democratic intellectual faith.”77 Dewey, for example, described personality as “the one thing of permanent and abiding worth” and a thing that lies “in every human individual.”78 Throughout the nineteenth century, personality was, in other words, widely understood as fundamental to and a function of humanity. When Justice Holmes called on “personality,” he did so as part of this “individualistic and egalitarian cultural tradition that glorified [it] and that did so in part because it was understood to adhere in every human being.”79
Personality was, then, never meant to be a challenge for copyright registrants. At the time, it simply put a name to a widely held intuition: that the capacity for originality is something latent in all people, that it’s just a part of being human. And therein lies the rub. For hundreds of years, attempts at registering works by purported non-human authors presented little challenge for the courts. The courts didn’t need personality — so it fell to the wayside.
Indeed, so short-lived was personality’s time in the spotlight that the story of its erasure begins only a few paragraphs after its introduction in Bleistein. In his Bleistein opinion, Justice Holmes laid out a new conception of progress for copyright law: an “essentially industrial view of the purpose of aesthetic production.”80 It was, in other words, a conception of progress based on market value.81 The consequences of Holmes’s framework were twofold. First, it “separated the basis of copyright rights, personality, from the purpose of granting those rights, progress in the form of ‘commercial value.’”82 And second, in so doing, it initiated American copyright law’s “devol[ution] into a commodity-oriented regime focused on aesthetic objects,” rather than “aesthetic subjects.”83 That is to say, personality and progress were separated and the latter prioritized to the point of expunging the former.84 Now, decades later, “commercial value” defines both the basis and purpose of copyright, and “personality” is “more or less forgotten or seen at best as having always been an empty or meaningless category.”85
But this “misreading of Bleistein”86 need not spell the end of personality in copyright. Justice Holmes’s invocation of personality was not intended — nor should it be understood — as a “retreat” from the complications of aesthetic judgment.87 “On the contrary, personality stood on its own in Bleistein . . . not as a last resort but as something of paramount value.”88 Indeed, in the Hegelian sense, “property is the first embodiment of freedom and so is in itself a substantive end.”89 That property becomes one’s own by way of “placing [one’s] will in [it]”90 — that is, by placing one’s personality in it — suggests that the paramount value of the creative process is no less than the pursuit of freedom: “the goal of the person.”91 Thus, while its “great virtue” was once its “broad egalitarian inclusiveness,”92 recovering personality as the standard of originality today may provide a type of exclusivity unnecessary (and technologically unimaginable) in the nineteenth century — but a type of exclusivity that reflects what copyright law was always meant to protect: human creativity.
III. Personality at Play
Hints of this personality-driven conception of originality can be disinterred from existing doctrine, even in cases more frequently lauded for their other contributions to copyright law. Indeed, “personality” is frequently at play beneath the surface of American copyright law, both in judicial opinions and in scholarly debates.93 This section looks at both those contexts, applying the language of personality to make sense of a tension rising in contemporary debates about copyright and generative AI: the relationship between moral rights and romantic authorship.
Feist Publications, Inc. v. Rural Telephone Service Co.94 is one of the Court’s seminal originality cases, in which it held that alphabetical listings in phone books were factual compilations “devoid of even the slightest trace of creativity.”95 The Court rested its conclusion that the plaintiff’s white pages lacked the requisite originality on its finding that the work was only a “mechanical” arrangement of otherwise uncopyrightable facts96 — that they were not expressive. The typical reading of Feist thus suggests that the decision falls squarely in line with the general pattern in originality cases: that the originality inquiry is little more than a restatement of copyright’s idea/expression distinction.97
But buried in the Feist decision are intimations of something more in originality. Consider, for example, the hypothetical that the Court offered to illustrate its proposition that originality does not require novelty.98 The Court described “two poets, each ignorant of the other, [who] compose identical poems.”99 In that case, the Court matter-of-factly explained that “[n]either work [would be] novel, yet both [would be] original and, hence, copyrightable.”100 Such an example does well to illustrate the concept of “independent creation”101 — that even identical works may be original so long as they were not produced by copying — but it does little to explain why such an outcome is normatively desirable. And without further explanation, it seems an unusual if not out-of-place addition to the rest of the Court’s explication of the originality requirement, which is otherwise grounded not in the Court’s own illustrations but in citations to existing authority.102 What, then, might be gleaned from its choice to include the two poets?
This Note posits that the Court — perhaps unwittingly — drew out originality’s rejection of a novelty requirement not because independent creation is particularly difficult to understand or especially worthy of emphasis but instead because it was voicing a value judgment, one grounded in the same intuition that Justice Holmes evoked in Bleistein. That is, the Court humanized these principles of copyright because they speak to the normative interests underpinning the law: When the Court said it didn’t care about the novelty of the two poems, it implied that what it did care about was that they came from the two poets. It’s a distinction worthy of illustration because it makes clear that originality is found not in works themselves but in their authors. That the Court imbued its explanation with personality — within the two poets — hints at an otherwise unspoken assumption: that while authorship may “presuppose” originality, originality itself presupposes personhood.
Now consider an alternative hypothetical. This time, there is one poet and one generative AI platform. Like in the Feist Court’s example, imagine that the poet writes a poem. At the same time, the AI is prompted to “write a poem” of its own. Though neither the poet nor the AI (nor, perhaps more aptly, its prompter) knew of the other, the poems produced are identical. Again, neither poem would be novel — but would the Feist Court maintain that both were original? Would it matter if the poet’s previous works had been included in the AI’s training data? Those who are concerned about generative AI’s impact on the creative arts are likely to bristle at the suggestion that this scenario warrants an outcome as cut-and-dry as the one imagined for the Feist Court’s two poets.
But if intuitions about the “right” outcome in the latter example differ from the Court’s dicta in Feist, it suggests that the poet hypothetical tells us something about originality beyond its disinterest in novelty. If all that is required of originality is independent creation and a bit of creativity,103 it’s hard to explain why the AI-generated poem would be any less deserving of the Court’s sign-off than its human-generated analog. Indeed, one scholar has questioned whether it’s possible to “rationally defend the proposition that copyright rewards originality and creativity and at the same time, treat human- and AI-generated works as legally distinct — even if we can’t tell the difference.”104
What the poet hypotheticals offer, then, is an opportunity to put a finger on “the difference” between human- and AI-generated works and, in doing so, to defend their difference in treatment. The difference, of course, seems almost too simple: In the Feist Court’s example, both poets were human; in the second, one was not. What these hypotheticals suggest, in other words, is that — if the two scenarios warrant different outcomes — what copyright cares about is not, at its most fundamental, novelty or creativity at all. Instead, what matters most is simply that a human has done the creating. That is to say, copyright assumes — first and foremost — that originality is a human characteristic.
Though it is a simple proposition, its exposition is useful not only because it offers a rational normative basis for differentiating between the treatment of human- and AI-generated works but also because it helps resolve some of the tensions at play in contemporary debates about the shifting contours of copyright protection. Consider, for example, the question of expanding moral rights. Translated from the French notion of droit moral, moral rights are “non-economic rights that are considered personal to an author,” like the right of attribution (“the right . . . to be credited as the author of their work”) and the right of integrity (“the right . . . to prevent prejudicial distortions of their work”).105 As the linguistic ancestry may suggest, moral rights feature more prominently in European copyright law106 and were formally adopted in the United States only in 1989.107 Still today, moral rights in the United States are relatively limited and extend only to “works of visual art” as defined by the Visual Artists Rights Act of 1990.108 Increasingly, however, commentators have proffered the possibility of more robust moral rights as a potential shield against threats to creative works, like generative AI.109
The problem with these suggestions is that, as some scholars have pointed out, it’s “ironic” to “import[] European notions of ‘moral rights’” while simultaneously “marshaling . . . criticism against” another fundamental principle of European copyright: the romantic author.110 Romantic authorship imagines a creative genius, someone whose “originality of spirit” translates into “originality of form.”111 It is a vision of authorship that would appear to suit a personality-based conception of originality quite nicely.112 But while much ink has been spilled over the impact of literary romanticism on American copyright law, many scholars reject the “romantic author” figure.113 And indeed, some have suggested that “Bleistein’s very low standard of originality [actually] disproves” that “romanticism had any real influence.”114 How romantic can our vision of genius be, after all, if anyone can meet the mark?
But Beebe offers a rebuttal. Rather than rejecting romantic authorship in toto, Bleistein, he says, actually “embraced a specifically American romantic conception of the author.”115 Unlike other “strains,” American romanticism does not imagine “any stereotyped notion of heroic daemonic genius.”116 Instead, the American romantic author is “distinctively democratic” — an “everyday, common genius.”117 And with this more “liberal, egalitarian, and humanistic”118 image of authorship in mind, Beebe suggests that Bleistein was “arguably the high point of American romanticism’s influence on our copyright law.”119
This alternative conception of romantic authorship offers a critical foothold for understanding contemporary debates about generative AI. The purported irony of rejecting romance while embracing moral rights is resolved if one recognizes that American copyright law isn’t interested in protecting the heroic genius.120 Criticisms levied against that notion of romantic authorship can exist in tandem with an interest in protecting the other. Beebe’s description of American romanticism speaks to copyright’s interest in protecting something more basic and more fundamental than what’s valorized by the European notion of romantic authorship: something more like creativity for creativity’s sake.121 The American conception of romanticism sees creativity worth protecting in that which is nothing more than human.122 This — human creativity — is what moral rights are meant to protect.123
Together, then, this vision of American romanticism, backed by personhood-based theories of property, and the Feist Court’s poet hypothetical tell us something about the core of copyright. The Court didn’t care whether the two poems were novel. It didn’t even ask whether they were “good” by any measure of aesthetic value. It needed only to know that two people had independently created them.124 That the poems came from poets was enough — the Court recognized the worth of their wills. What American copyright law seeks to protect is as basic as that: It is an interest in human creativity. And this is what makes generative AI a misfit for our current regime. While it may be difficult to argue that generative AI programs are somehow incapable of originality as a matter of creative output, it’s an altogether easier task to assert plainly that AI-generated works are missing the one thing copyright has always been meant to protect. That is, they lack personality. Recovering personality as the premise of originality thus offers what copyright otherwise seems to be missing: a sure-footed normative distinction between human- and AI-generated works — and one that supports the adoption of more robust moral rights for human creators. Indeed, moral rights — themselves an acknowledgment of human rights125 — that preserve copyrights for human-generated works should flow naturally from the recovery of personality in American copyright law.126
IV. The Weight of Personality
As stakeholders seek to revamp copyright to meet the challenges of generative AI, recovering personality as the basis of originality can provide benefits distinct from the outcomes achieved by the human authorship requirement. Indeed, though originality and authorship have been “intertwined” “throughout U.S. copyright law,”127 disentangling them shows that they serve entirely distinct ends. While authorship makes copyright functional, originality makes copyright meaningful, and recovering that purpose — the recognition of personality — allows authors to speak to personality interests, too. More so than the language of economic injury, the language of personality allows authors to better give voice to the reality of their own perceived harms in the face of generative AI. In doing so, recognizing personality as the foundation of originality and, thus, of copyright also substantiates the conclusion that copyright owners should be compensated for the use of their works by generative AI.
Authorship is about incentives. Indeed, “Authors” — unlike originality — are mentioned in the Copyright Clause as the subject of the Constitution’s commitment “[t]o promote the Progress of Science and useful Arts.”128 To achieve this end, the Framers offered authors “the exclusive Right to their respective Writings.”129 This incentive-based theory of copyright “posits a straightforward exchange between artists and society.”130 Copyright offers authors “temporally granted bounded monopolies” in exchange for “original works that enrich the cultural landscape.”131 Without these incentives, the theory presumes that artists would choose not to create at all, because “uncompensated third parties could freely copy and distribute their works.”132 Thus, while “the public” may be the “primary intended beneficiary” of copyright,133 the identification and subsequent incentivization of authors is necessary to see that benefit enjoyed.
The problem is that thinking about copyright as an incentive structure can only get you so far, particularly in ever-changing and ever-more challenging technological landscapes. Today, it may explain the human authorship requirement — non-humans will not be motivated by the benefits of copyright protection — but if “[t]he goal of American copyright law” is not to “reward authors” but “to incentivize the creation of works that benefit the public,” then “the question of whether a computer can be an author [should be] irrelevant,” so long as it is producing “marketable works.”134 Recognizing this tension, Professor Zachary Catanzaro argues that “[i]f human authorship is to remain the focus of American copyright law, it can no longer do so based solely on previous conceptions of originality and the strength of the incentive rationale.”135
But while Catanzaro suggests “shift[ing]” focus “toward strengthening the moral rights of human authors,”136 this Note offers an alternative — or at least a preliminary — tack: reviving a more robust conception of originality. As Catanzaro puts it, “If Congress believes that human art deserves more protection than machine outputs, that there is something unique about human intentionality in creating new expression, then it should say so.”137 All this Note proposes is that, before courts turn to new frameworks, they say what has long been taken for granted: that at the core of our discussions about creativity and authorship and moral rights is an understanding that originality is, at its core, about humanness — about personality.
This acknowledgment speaks to the independent purpose of the originality requirement. Authorship is functional in that it answers the question of “who” copyrights belong to, and restrictions on authorship reflect theories of “who” copyright’s incentives will motivate. It’s logical, then, that “Congress has traditionally viewed authorship as a policy question, not a metaphysical or teleological one.”138 Originality, by contrast, answers “what” copyright is meant to protect, and restrictions (or, as explored above, lack of restrictions) on originality reflect what types of creative content copyright values: that is, those rooted in personality.
That protecting personality is the telos of copyright speaks to a long-standing cultural understanding about the value of art and creativity. Indeed, eighteenth-century “commentators believed that the progress of the fine arts promised to promote the overall progress of civic virtue and good government,”139 and the Framers themselves “clearly subscribed” to the idea “that ‘a flourishing state of the Arts and Sciences[] contributes to National prosperity and reputation.’”140 Today, too, commentators maintain that “[a]rt is not something that happens at the periphery of our lives” but is instead one of “the most important vehicles by which we come to understand one another.”141 It “helps shape our societ[y] by providing platforms for creativity and solidarity to thrive.”142 And it “matters today more than ever because it outlives the contentious political veneer that is cast over everything.”143 Critically, copyright recognizes that these public benefits come not from the state itself but from individuals — the law exists to support the creativity of the people. Put differently, the “[a]rt matters because artists matter.”144 When parties claim that generative AI is a threat to the arts, these are the stakes. It’s not just about art, so to speak. It’s about protecting personality.
Recovering personality as the basis of originality thus offers a hook for litigants to tell a more accurate story about the harms of generative AI. To illustrate, compare two sets of cases brought by one plaintiff: the Authors Guild. “[T]he Authors Guild is the nation’s oldest and largest professional organization for published writers” and includes in its mission “advocat[ing] for the rights of writers,” including copyrights, which it sees as “critical to a healthy literary culture.”145 Students of copyright likely recognize the Guild from the “Google Books case”146 and the parallel suit against HathiTrust, the library collective that digitized books to create a searchable database and improve accessibility for print-disabled readers.147 In those cases, the Guild argued that the mass compilation and digitization of copyrighted works was infringement under 17 U.S.C. § 106 — but in both cases, the Second Circuit held that the defendants’ works were protected as fair use.148
The fair use analysis involves consideration of four factors,149 and the fourth asks specifically about “the effect of the use upon the potential market for or value of the copyrighted work.”150 The Supreme Court has dubbed the fourth factor “the single most important element of fair use”151 — a superlative the Google court saw as “consistent with the fact that the copyright is a commercial right.”152 Unsurprisingly, then, both the Google and the HathiTrust cases paid attention to the commercial impact of the defendants’ acts.153 And when the Authors Guild voiced its claims to the court, it focused on its economic injuries.154
But while the Second Circuit’s fair use analyses posed questions distinct from those asked in the Bleistein Court’s copyrightability inquiry, it’s notable that each case featured commerciality. Indeed, if one accepts Beebe’s proposition that, post-Bleistein, commercial value overtook the other interests underlying originality,155 then perhaps it’s true that other interests underlying copyright, generally, are similarly overlooked when commercial impact dominates copyright litigation. That is, copyright litigation is spoken in the language of profits and livelihoods; it is, on various axes, a market-based harm analysis. But the creative arts have never been concerned only with economics. Confining litigants to this vocabulary prevents them from voicing the totality of their felt harms. Examining the Authors Guild’s more recent case makes this clear.
Last fall, the Guild filed another class action complaint — this one against OpenAI. The Guild alleged that, like Google and HathiTrust, OpenAI had infringed the plaintiffs’ exclusive rights by using their copyrighted works to train its large language AI model.156 As it had before, the Authors Guild described the harm incurred largely in terms of lost income and diminished job opportunities.157 But unlike in its complaints against Google and HathiTrust, the Guild also voiced its concern that “OpenAI’s conduct poses [risks]” not only to the livelihoods of the “[p]laintiffs and other professional writers” but also to “the literary arts generally.”158 And though this variety of harm was mentioned only the once, it matters because it better reflects what the Guild has in other contexts expressed as the source of its discontent with generative AI.
Indeed, we know a great deal about what the Guild’s AI harm story sounds like when it’s unconstrained by legal briefing standards. In other forums, the story the Guild tells is not so much about the incentives necessary for creation or the purely economic consequences of generative AI. It is, instead, a story about the value of human creativity itself.
Take, for example, comments that the Guild submitted in response to the Copyright Office’s inquiry on the impact of AI technologies on copyright. In those remarks, the Guild forthrightly stated that AI “force[s] us to interrogate our beliefs about human nature and art” and that “[a] world where machines in the form of AI dominate our arts is not a world most of us will want to live in.”159 It pleaded to the Copyright Office that arts are “fundamentally important to every human society” because they “help us feel and see . . . truths about human existence.”160 AI-generated works — “[w]hile [they] might look or sound like human-created works” — will inevitably “lack[] the essential human faculties that move the arts forward.”161 They “will never add that spark of human creativity — the elements that make[] art true art.”162
And, in fact, the Guild has been explicit that these concerns — about the value of art and human creativity — precede its commercial concerns, framing its copyright interests through the lens of cultural value. It explained “that it is crucial for our culture and the future of democracy to ensure that our literary arts remain vibrant and diverse”163 and cautioned that “[b]eyond the economic impact on writers, the unregulated development and use of generative AI technologies will lower the quality of books, journalism, and public discourse fundamental to democratic culture.”164 And — harkening back to Beebe’s description of the post-Bleistein collapse between personality and commercial progress — the Guild warned that “[i]n the debates around the commercial uses of copyright, we tend to forget copyright’s role in shaping our democracy as the engine of self-determination, ideas, and expression.”165
This is the Authors Guild’s story about generative AI. Unlike in its previous litigation, the reality of the Guild’s perceived harms is more than can be conveyed through only claims of lost profits. Its story is about harms to something simultaneously more fundamental and more esoteric than market value can bespeak: the social, political, and cultural value of creators, artists, and “writers whose works spring from their own minds and their creative . . . expression.”166 The Guild’s story is an invocation of personality. But it’s left to wither on the vine if plaintiffs are compelled to tell the story of their injuries with only the sterile and restrictive language of economic harm.
In its decision in Authors Guild v. Google, Inc.,167 the Second Circuit itself acknowledged that “[t]he ultimate goal of copyright is to expand public knowledge and understanding” and that “the ultimate, primary intended beneficiary is the public, whose access to knowledge copyright seeks to advance.”168 Recovering personality as a first principle of copyright allows litigants — like the Authors Guild — to speak more explicitly about these ends, to avoid filtering their stories through an artificially restrictive sieve between cultural and economic value, and to remind courts exactly what it is that copyright is meant to protect.169 In so doing, these stories beg the conclusion that the value chain created by generative AI must, eventually, result in benefits to the human creators who made the technological project possible. The invocation of personality, in other words, justifies economic payout to human creators.
Such a conclusion is supported by the personhood theories of property. As Radin explains, the relationship between property and personhood exists on a continuum.170 On one end of the spectrum are fungible things, like money.171 On the other are personal things, defined by the “character or strength of the connection” between the thing and one’s personhood.172 This spectrum view “generates a hierarchy of entitlements” that privileges rights over things closer to the “personal” end and deprioritizes rights over the fungible.173 Indeed, “fungible property rights . . . can be overridden in some cases” where “personal property rights . . . cannot be.”174 Understanding where the variable “relationships between persons and things . . . fall . . . [on] this continuum” offers “a guide to determine which property is worthier of protection.”175 In other words, the personhood perspective offers us “an explicit source of values for making moral distinctions in property disputes.”176
Recognizing that copyright assumes personhood — in the language of “personality” — as necessary for originality means only that copyrighted works will always fall closer to the personal side of the spectrum. “[P]ayments from intellectual property users to the property creator are [thus] acts of recognition[,] . . . and it is through such acknowledgment that an individual is recognized by others as a person.”177 In the face of generative AI, this kind of recognition is paramount. As exemplified in the story of the Authors Guild, fears about generative AI are not merely fears about economic loss — the kind of loss one might feel after the theft of a fungible good.178 They are fears “specially related to personhood”179 and should be recognized and compensated as such.
Conclusion
Today, recovering the language of personality as the basis of copyright’s originality inquiry offers people the opportunity to more accurately and authentically give voice to their concerns about generative AI. As authors, artists, and creators try to protect their creative works from the challenges brought by this technology, the language of market harm that now dominates copyright litigation is insufficient. Understanding how copyright should respond to generative AI requires reckoning with what copyright should care about and what it should protect. The history of originality and the language with which we talk — and historically have talked — about creativity suggests that the answer to these questions has less to do with the economic value of any particular work than with the inherent value of the individuals who create them.