Contributions

Expansion of the Copyright Act: The Inclusion of Art Created In-Part by an Artificial Intelligence Program

by Nao­mi Per­la1

The pur­pose of the intel­lec­tu­al prop­er­ty regime, as set out by the Intel­lec­tu­al Prop­er­ty Clause, is to  incen­tivize artists and inven­tors to cre­ate and pro­duce works that will ben­e­fit soci­ety.2 While the Copy­right Act aims to fur­ther these goals by grant­i­ng pro­tec­tion for var­i­ous works, Con­gress also intend­ed to cre­ate some lim­i­ta­tions as to the types of works afford­ed pro­tec­tion.3 Accord­ing­ly, the Copy­right Act stip­u­lates that a work must be con­sid­ered an “orig­i­nal work of author­ship” pur­suant to 17 U.S.C. § 102(a).4 The main pur­pose of this state­ment is to set out the orig­i­nal­i­ty require­ment con­sti­tu­tion­al­ly nec­es­sary to estab­lish copy­right pro­tec­tion.5 The require­ment of “author­ship,” although not explic­it­ly artic­u­lat­ed by courts as a con­sti­tu­tion­al require­ment dis­tinct from orig­i­nal­i­ty, is an imper­a­tive com­po­nent of the analy­sis that courts have grap­pled with in unique cir­cum­stances.6 The issue of author­ship becomes increas­ing­ly sig­nif­i­cant as tech­nol­o­gy advances and replaces the human author in some capac­i­ty, there­by forc­ing courts to rein­ter­pret the mean­ing of “author­ship” with­in the Copy­right Act.7

The uti­liza­tion of tech­nol­o­gy specif­i­cal­ly with­in the art space pos­es chal­lenges for courts faced with deter­min­ing whether a work of art can be con­sid­ered an “orig­i­nal work of author­ship.”8 To sat­is­fy the orig­i­nal­i­ty require­ment, an author must inde­pen­dent­ly cre­ate the work and pro­vide a “mod­icum of cre­ativ­i­ty.”9 More­over, accord­ing to the Com­pendi­um of U.S. Copy­right Office Prac­tices (“Com­pendi­um”), “to qual­i­fy as a work of ‘author­ship’ a work must be cre­at­ed by a human being.”10 Works of art that are cre­at­ed by a tech­no­log­i­cal mech­a­nism that uses ran­dom mechan­i­cal process­es and requires no human inter­ven­tion fail to sat­is­fy the Copy­right Act’s require­ments, there­by falling into the pub­lic domain.11 When artists uti­lize tech­nol­o­gy to assist them in cre­at­ing their art pieces, they arguably delete them­selves from a part of the cre­ative process. This rais­es the ques­tion of whether these art pieces are authored by a human, and whether there is suf­fi­cient cre­ativ­i­ty present to sat­is­fy the orig­i­nal­i­ty require­ment. More recent­ly, artists have start­ed to use arti­fi­cial intel­li­gence (“AI”) pro­grams to cre­ate their art, which has exac­er­bat­ed these exist­ing con­cerns about author­ship and orig­i­nal­i­ty.12 AI pro­grams per­form con­sid­er­ably more of the work than a stan­dard tech­no­log­i­cal mech­a­nism, which in turn leads to less cre­ativ­i­ty being pro­vid­ed by the human author. As such, it is unclear whether art work cre­at­ed using AI pro­grams are wor­thy of copy­right pro­tec­tion. 

This Con­tri­bu­tion argues that art pro­duced by a human cre­ator with the assis­tance of an AI pro­gram has been “authored” to the extent required by the Copy­right Act, and sat­is­fies the Act’s orig­i­nal­i­ty require­ment. Accord­ing­ly, such art falls with­in the def­i­n­i­tion of an “orig­i­nal works of author­ship” pur­suant to 17 U.S.C. § 102(a), and is wor­thy of copy­right protection.

*****

The phrase an “orig­i­nal work of author­ship” can be parsed out into two com­po­nents that must be sat­is­fied: author­ship and orig­i­nal­i­ty.13 The “orig­i­nal­i­ty” com­po­nent has been clear­ly artic­u­lat­ed by the Supreme Court, while “author­ship” has been left some­what unat­tend­ed.14

Although the Copy­right Act does not pro­vide a def­i­n­i­tion of “author­ship,” courts have inter­pret­ed the phrase broad­ly and with enough flex­i­bil­i­ty to reflect the ever­chang­ing cre­ative land­scape.15 That said, the def­i­n­i­tion of “author­ship” is not with­out lim­its.16 Even the Ninth Cir­cuit, which argued for an expan­sive def­i­n­i­tion of “author­ship” in Gar­cia v. Google, lat­er sought to lim­it the term’s scope in Naru­to v. Slater.17 In Naru­to, the Ninth Cir­cuit explained that a flex­i­ble def­i­n­i­tion of “author­ship” did not mean that the term had no bounds.18 Accord­ing­ly, the court denied stand­ing under the Copy­right Act to a mon­key who took a self­ie because “author­ship” did not include ani­mals.19 Sim­i­lar­ly, the Com­pendi­um defines “author­ship” as requir­ing a human cre­ator; there­fore, when there is absolute­ly no human inter­ven­tion, author­ship is not present.20

The com­pro­mise between these two approach­es lies between requir­ing a human cre­ator and rec­og­niz­ing author­ship in sit­u­a­tions where the human is con­tribut­ing less than what is cus­tom­ary. For instance, courts have grant­ed author­ship to human cre­ators and con­sid­ered the orig­i­nal­i­ty require­ment to be sat­is­fied in cas­es where the human creator’s involve­ment was less than tra­di­tion­al­ly expect­ed.21 For exam­ple, in Bur­row Giles v. Sarony, the Supreme Court grant­ed author­ship to a pho­tog­ra­ph­er, even while rec­og­niz­ing that it is the cam­era that cap­tures the image and trans­forms it into a medi­um that can be show­cased to the pub­lic.22

Although the two com­po­nents are sep­a­rate, they are relat­ed in that deter­min­ing whether the human cre­ator is the author can be depen­dent on whether the cre­ator has con­tributed the “mod­icum of cre­ativ­i­ty” nec­es­sary to sat­is­fy the orig­i­nal­i­ty require­ment.23  In Feist Pub­li­ca­tions Inc. v. Rur­al Tel. Serv., the Supreme Court empha­sized that the bar for orig­i­nal­i­ty is low.24 The Court fur­ther held that if cre­ativ­i­ty could not be found in the work itself, a court may look to the man­ner in which the author select­ed, com­piled, or arranged var­i­ous ele­ments to sat­is­fy the require­ment.25 Years pri­or to Feist, the Court uti­lized this form of rea­son­ing and applied it to pho­tographs in Bur­row-Giles, con­clud­ing that the photographer’s cre­ative con­tri­bu­tions in select­ing, com­pil­ing, and arrang­ing the var­i­ous ele­ments in the pho­to were suf­fi­cient to sat­is­fy the orig­i­nal­i­ty require­ment, despite the large per­cent­age of work the cam­era was con­tribut­ing, and grant­i­ng author­ship to the pho­tog­ra­ph­er and copy­right pro­tec­tion over the pho­to.26

In deter­min­ing whether ade­quate orig­i­nal­i­ty is being pro­vid­ed by the human cre­ator, some courts have designed unique tests for par­tic­u­lar cir­cum­stances that help sat­is­fy the orig­i­nal prin­ci­pals laid out in Feist: the lion’s share bal­anc­ing test27 and the suf­fi­cient con­trol test.28 The bal­anc­ing test is specif­i­cal­ly applied when a soft­ware pro­gram is uti­lized.29 This test com­pares the program’s con­tri­bu­tions in cre­at­ing the art work to the con­tri­bu­tions of the pro­gram user. If the program’s work con­sti­tutes the lion’s share of the work and the user’s con­tri­bu­tions were mere­ly mar­gin­al, then the pro­gram­mer would be able to gain copy­right pro­tec­tion over the out­put and vice ver­sa.30

The con­trol analy­sis test looks at whether the human cre­ator has suf­fi­cient con­trol over the cre­ative process, even if oth­er enti­ties con­tributed cre­ative­ly as well.31 For instance, in Lind­say v. Wrecked & Aban­doned Ves­sel R.M.S. Titan­ic,  the court found that the direc­tor held the copy­right over the film regard­less of the var­i­ous con­tri­bu­tions pro­vid­ed by oth­er indi­vid­u­als, since he was mas­ter­mind­ing the entire process, there­by ensur­ing his orig­i­nal ideas were being pro­duced.32 The Third and Fifth Cir­cuits applied this analy­sis where tech­nol­o­gy was employed, and both courts found in favor of the indi­vid­ual claim­ing author­ship and pro­tec­tion.33

Although courts have applied the analy­ses above and eval­u­at­ed var­i­ous forms of art cre­at­ed with tech­nol­o­gy, art cre­at­ed with the assis­tance of an AI pro­gram has large­ly been left unevaluated.

*****

The use of arti­fi­cial intel­li­gence pro­grams to assist in cre­at­ing works of art is a nov­el and unique process, and the Copy­right Act should expand to include these art works with­in its bor­ders. Pro­vid­ing pro­tec­tion for such art is con­sis­tent with the his­to­ry of the Copy­right Act, which is one of grad­ual expan­sion,34 and the way in which the Act is writ­ten.35  Con­gress inten­tion­al­ly left “author­ship” unde­fined.36 Rather than attempt­ing to pre­dict the ways in which art would evolve, Con­gress rec­og­nized it would bet­ter suit the copy­right regime to leave terms ambigu­ous, which would allow the Act to adapt to the times and keep the copy­right regime from ceas­ing to exist.37 The Supreme Court acknowl­edged the Act’s expan­sive­ness in Bur­row Giles, where the cam­era was rec­og­nized as mere­ly a tool uti­lized by artists to cre­ate their works.38 The Court opined that “the only rea­son why pho­tographs were not includ­ed in the extend­ed list in the act of 1802 is prob­a­bly that they did not exist, as pho­tog­ra­phy as an art was then unknown .…”39 This analy­sis can be applied to AI pro­grams, which, like cam­eras, are mere­ly uti­lized by artists as a tool to cre­ate. Arti­fi­cial intel­li­gence is sim­ply the next step in the advance­ment of technology.

Fur­ther­more, courts can eas­i­ly apply the orig­i­nal­i­ty analy­sis to the AI con­text by deter­min­ing whether the con­tri­bu­tions by the artist are suf­fi­cient to sat­is­fy the orig­i­nal­i­ty require­ment. When an artist oper­ates an AI pro­gram, they must first deter­mine the input that the AI pro­gram will use to gen­er­ate the out­put.40 The AI pro­gram can only cre­ate based on the spe­cif­ic rules pro­vid­ed by the input, which in turn has been pro­vid­ed by the artist.41 Addi­tion­al­ly, an artist using an AI pro­gram must also choose which of the out­put will be show­cased to the pub­lic and in what medi­um they will be viewed in.42 For instance, an artist may choose spe­cif­ic art pieces that have par­tic­u­lar char­ac­ter­is­tics, like chil­dren between the ages of one and five, as the input. The AI pro­gram will cre­ate the out­put based on that set of rules pro­vid­ed by the artist. Once  com­plet­ed, the artist will look at all the works cre­at­ed and per­haps select five to be con­vert­ed to 3‑D can­vas for opti­mal view­ing. All of these choic­es alone should sat­is­fy the low bar of the orig­i­nal­i­ty require­ment. How­ev­er, if these choic­es as a whole are insuf­fi­cient, orig­i­nal­i­ty can be sat­is­fied by the mere selec­tion and arrange­ment of the ele­ments, as per Feist.43 Because the artist using an AI pro­gram typ­i­cal­ly selects and com­piles the input, selects the out­put, and arranges the final prod­uct, the orig­i­nal­i­ty require­ment is fulfilled.

Although the analy­sis above seems to sat­is­fy the require­ments, the Sev­enth Circuit’s deci­sion in Kel­ley v. Chica­go Park Dis­trict rais­es some con­cerns. In Kel­ley, the Sev­enth Cir­cuit con­clud­ed that the author­ship and orig­i­nal­i­ty require­ments were not met where a park exhib­it was cre­at­ed by a human cre­ator using flower arrange­ments.44 Even though the human cre­ator select­ed the flow­ers and fash­ioned the ini­tial arrange­ment, the court opined that the gar­den would evolve based on the forces of nature, and would con­tin­ue to change long after the cre­ator com­plet­ed his work.45 This ratio­nale may prove prob­lem­at­ic in the AI con­text because once the cre­ator plugs in the data set, the AI pro­gram is essen­tial­ly left to its own devices. How­ev­er, there is a key dis­tinc­tion: the AI pro­gram can only func­tion based on the pre­de­ter­mined rules set out by the input pro­vid­ed by the cre­ator. In con­trast, the artist in Kel­ley does not have con­trol over nature, even if he ini­tial­ly picks the flow­ers. Thus, unlike a gar­den, an AI pro­gram must abide by the guide­lines set for it.

Because AI is such a nov­el con­cept, the lion’s share bal­anc­ing test, which was cre­at­ed pre­cise­ly to deal with tech­no­log­i­cal pro­grams, may be a bet­ter method for deter­min­ing whether the orig­i­nal­i­ty require­ment is sat­is­fied.46 For instance, in Rear­den LLC, the court deter­mined that the actions tak­en by the actors and direc­tors could not be con­sid­ered mere­ly mar­gin­al since their cre­ative choic­es were large­ly reflect­ed in the fin­ished prod­uct.47 Accord­ing­ly, the pro­gram­mer could not claim that the program’s con­tri­bu­tions con­sti­tut­ed the lion’s share of the work.48 When an artist uti­lizes an AI pro­gram, their cre­ative choic­es are man­i­fest­ed in the fin­ished art pieces as well. There­fore, when com­par­ing the work of the AI pro­gram to that of the artist, the artist’s con­tri­bu­tions are far from mar­gin­al and rep­re­sent the lion’s share of the work.49

The last method that could be applied by courts to deter­mine orig­i­nal­i­ty is the suf­fi­cient con­trol test.50 An artist who choos­es the input, which cre­ates the rules the AI pro­gram must fol­low, and mod­i­fies the out­put, demon­strates a high degree of con­trol over the process and the fin­ished prod­uct, there­by ful­fill­ing the orig­i­nal­i­ty require­ment, as per Lind­say.51 All of these deci­sions indi­cate that the artist is the one mas­ter­mind­ing the process to bring their cre­ative con­cepts to life.52 How­ev­er, irre­spec­tive of these fac­tors, accord­ing to the Supreme Court in Com­mu­ni­ty for Cre­ative Non-Vio­lence v. Reid, the author must be the one who trans­lates the idea into the fixed medi­um of expres­sion.53 The appli­ca­tion of this con­cept to the AI con­text poten­tial­ly leads to some com­pli­ca­tions, since it is the AI pro­gram that takes the ideas set forth in the input and express­es them through the out­put. How­ev­er, this hold­ing did not pro­scribe pro­tec­tion in many oth­er instances where tech­nol­o­gy was uti­lized in a sim­i­lar fash­ion, and should not pro­hib­it pro­tec­tion in the AI con­text either.54 No mat­ter which lens one views the orig­i­nal­i­ty require­ment, the fin­ished prod­uct reflects a myr­i­ad of cre­ative choic­es, and that is suf­fi­cient to sat­is­fy the “mod­icum of cre­ativ­i­ty” nec­es­sary to ful­fill the requirement.

Final­ly, the Con­sti­tu­tion clear­ly states that the pur­pose of the Intel­lec­tu­al Prop­er­ty Clause is to pro­mote the progress of the arts by grant­i­ng rights to authors for a lim­it­ed peri­od of time.55 This pro­duces an eco­nom­ic incen­tive that moti­vates artists to cre­ate and dis­sem­i­nate their works to the pub­lic. With­out pro­tec­tion, artists will lose this impor­tant incen­tive.56 As the Supreme Court opined in Twen­ti­eth Cen­tu­ry Music Corp. v. Aiken, “when tech­no­log­i­cal change has ren­dered its lit­er­al terms ambigu­ous, the Copy­right Act must be con­strued in light of this basic pur­pose.”57 If copy­right pro­tec­tion is denied to artists who cre­ate art with the assis­tance of an AI pro­gram, then the indus­try will dimin­ish, leav­ing a gap in the econ­o­my where those works of art belong and depriv­ing the pub­lic of their ben­e­fit.58

*****

Con­gress intend­ed for the Copy­right Act to expand and accom­mo­date tech­no­log­i­cal and artis­tic advances.59 Nev­er­the­less, a piece of art must sat­is­fy cer­tain requirements—namely the orig­i­nal­i­ty and author­ship requirements—in order to gain pro­tec­tion.60 Art pieces that were cre­at­ed by a human artist, with the assis­tance of an arti­fi­cial intel­li­gence pro­gram, are gen­er­al­ly able to sat­is­fy these require­ments, and should be includ­ed with­in the def­i­n­i­tion of an “orig­i­nal work of author­ship.” Courts tend to inter­pret “author­ship” broad­ly and have con­clud­ed that a human cre­ator can be con­sid­ered to have authored an art work even when their cre­ative choic­es are minor, there­by sat­is­fy­ing the orig­i­nal­i­ty require­ment.61 More­over, pro­vid­ing pro­tec­tion helps incen­tivize artists to con­tin­ue cre­at­ing, which sat­is­fies the pol­i­cy ratio­nale under­ly­ing the Copy­right Act.62

1. Nao­mi Per­la is J.D. Can­di­date (2021) at New York Uni­ver­si­ty School of Law. This piece is a com­men­tary on the prob­lem pre­sent­ed at the 2020 BMI (Broad­cast Music, Inc.) Moot Court Com­pe­ti­tion at Car­do­zo School of Law. The ques­tion pre­sent­ed asked whether works that have been cre­at­ed by a human artist and in part by arti­fi­cial intel­li­gence could fall with­in the def­i­n­i­tion of an “orig­i­nal work of author­ship” pur­suant to 17 U.S.C. § 102(a), there­by caus­ing the work to be wor­thy of copy­right pro­tec­tion. This Con­tri­bu­tion presents a dis­til­la­tion of the side of the argu­ment assigned to the author in the BMI com­pe­ti­tion, and the views expressed here­in do not nec­es­sar­i­ly reflect the views of the author.

2.  See U.S. Con­st. art. I, § 8, cl. 8; see also Sony Corp. of Am. v. Uni­ver­sal City Stu­dios, Inc., 464 U.S. 417, 429 (1984) (describ­ing the pur­pose of the lim­it­ed rights grant­ed to cre­ators by the Intel­lec­tu­al Prop­er­ty Clause to be to incen­tivize fur­ther creativity).

3. See Melville B. Nim­mer & David Nim­mer, 1 Nim­mer on Copy­right § 2.03 (Matthew Ben­der, rev. ed. 2019) (“A flex­i­ble def­i­n­i­tion was intend­ed that would nei­ther ‘freeze the scope of copy­rightable sub­ject mat­ter at the present stage of com­mu­ni­ca­tions tech­nol­o­gy [n]or … allow unlim­it­ed expan­sion into areas com­plete­ly out­side the present con­gres­sion­al intent.’” (inter­nal cita­tions omitted)).

4. 17 U.S.C. § 102(a).

5. See Feist Publ’ns, Inc. v. Rur­al Tel. Serv. Co., 499 U.S. 340, 348 (1991) (not­ing that the “orig­i­nal work of author­ship” com­po­nent of the Copy­right Act sets out the orig­i­nal­i­ty requirement).

6. See Andrien v. South­ern Ocean Cnty. Cham­ber of Com­merce, 927 F.2d 132, 135 (3d Cir. 1991) (find­ing author­ship for an indi­vid­ual who cre­at­ed a map by uti­liz­ing a print­er); Lake­dreams v. Tay­lor, 932 F.2d 1103, 1108 (5th Cir. 1991) (find­ing copy­right pro­tec­tion for t‑shirt designs cre­at­ed by humans and trans­posed by print­ers onto silkscreen);  Uran­tia Found. v. Maa­her­ra, 114 F.3d 955, 958 (9th Cir. 1997) (hold­ing a reli­gious book, even if cre­at­ed in part by a celes­tial being, could receive a copy­right for the input pro­vid­ed by the human author); Pen­guin Books U.S.A., Inc. v. New Chris­t­ian Church of Full Endeav­or, Ltd., 2000 U.S. Dist. LEXIS 10394, at *41 (S.D.N.Y. July 21, 2000) (“As a mat­ter of law, it is irrel­e­vant for copy­right pur­pos­es whether Jesus wrote the Course.”); Naru­to v. Slater, No. 15-cv-04324-WHO, 2016 U.S. Dist. LEXIS 11041, at *9–10 (N.D. Cal. 2016) (hold­ing that a mon­key can­not claim author­ship over a pho­to because the author­ship require­ment can only be sat­is­fied with a human cre­ator). 

7. See Bur­row-Giles Lith­o­graph­ic Co. v. Sarony, 111 U.S. 53, 56–60 (1884) (hold­ing a pho­tog­ra­ph­er as the author of his photographs).

8. Id.

9. See Feist Publ’ns, 499 U.S. at 346 (1991).

10. U.S. Copy­right Office, Com­pendi­um of U.S. Copy­right Office Prac­tices § 313.2 (3d ed. 2017).

11. Id. (“[T]o qual­i­fy as a work of ‘author­ship’ a work must be cre­at­ed by a human being.”).

12. See, e.g., Obvi­ous is a col­lec­tive of friends, artists and researchers, obvi­ous, https://obvious-art.com/ (last vis­it­ed Feb. 8, 2021).

13. 17 U.S.C. § 102(a).

14. See Feist Publ’ns, Inc., 499 U.S. at 347 (dis­cussing the orig­i­nal­i­ty com­po­nent of the phrase “orig­i­nal work of author­ship” at length, while leav­ing the author­ship com­po­nent less explored).

15. See Gar­cia v. Google, Inc., 786 F.3d 733, 741 (9th Cir. 2015) (“The statute pur­pose­ful­ly left ‘works of author­ship’ unde­fined to pro­vide for some flex­i­bil­i­ty[.]” (cit­ing Melville B. Nim­mer & David Nim­mer, 1 Nim­mer on Copy­right § 2.03 (Matthew Ben­der, rev. ed. 2019))).

16. Nim­mer on Copy­right § 102 (Matthew Ben­der, rev. ed. 2019); U.S. Copy­right Office, Com­pendi­um of U.S. Copy­right Office Prac­tices § 313.2 (3d ed. 2017) (stat­ing that to “qual­i­fy as a work of author­ship a work must be cre­at­ed by a human being”).

17. See Gar­cia, 786 F.3d at 741; see also Naru­to, No. 15-cv-04324-WHO, 2016 U.S. Dist. LEXIS 11041, at *9–10 (hold­ing that a pho­to tak­en by a mon­key could not gain copy­right pro­tec­tion because it was not cre­at­ed by a human).

18. Id.

19. Id.

20. Com­pendi­um (Third) § 306 (explain­ing that to “qual­i­fy as a work of author­ship a work must be cre­at­ed by a human being”).

21. See Andrien v. South­ern Ocean Cnty. Cham­ber of Com­merce, 927 F.2d 132, 135 (3d Cir. 1991) (find­ing author­ship for an indi­vid­ual who cre­at­ed a map using a print­er, where the print­er did most of the cre­ative work instead of the human); see alsoLake­dreams v. Tay­lor, 932 F.2d 1103, 1108 (5th Cir. 1991) (find­ing copy­right pro­tec­tion for t‑shirt designs cre­at­ed by humans, but where the print­ers did a sub­stan­tial amount of the work by trans­pos­ing the designs onto silkscreen).

22. See Bur­row-Giles Lith­o­graph­ic Co. v. Sarony, 111 U.S. 53, 56–60 (1884) (grant­i­ng author­ship to the pho­tog­ra­ph­er because of the cre­ative work he con­tributed in select­ing and arrang­ing the items with­in the photograph).

23. Id.

24. See Feist Publ’ns, 499 U.S. at 346 (not­ing that “even a slight amount [of orig­i­nal­i­ty] would suffice”).

25. Id. (dis­cussing the abil­i­ty for one to gain copy­right pro­tec­tion through the mere selec­tion, arrange­ment, or com­pi­la­tion of var­i­ous elements).

26. See Bur­row-Giles Lith­o­graph­ic Co., 111 U.S. at 56–60.

27. See Torah Soft Ltd. v. Dros­nin, 136 F. Supp. 2d 276, 283 (S.D.N.Y. 2001) (con­clud­ing that because defen­dant user only con­tributed a mar­gin­al amount of the work by inputting a sin­gle word into the pro­gram, no copy­right pro­tec­tion should be pro­vid­ed): Rear­den LLC v. Walt Dis­ney Co., 293 F. Supp. 3d 963 (N.D. Cal. 2018) (hold­ing no copy­right for the pro­gram­mer on the out­put cre­at­ed by the soft­ware pro­gram since the input con­tributed by the user was not mere­ly mar­gin­al); Dig­i­tal Drilling Data Sys. LLC v. Petrolink Servs., No. 4:15-CV-02172, 2018 U.S. Dist. LEXIS 83158, at *20–21 (S.D. Tex. May 16, 2018) (pro­vid­ing copy­right pro­tec­tion for the plain­tiff, the pro­gram cre­ator, over the schema which is cre­at­ed by the pro­gram, since the ele­ments of the schema already exist­ed in the pro­gram and was con­sid­ered the lion’s share of the work).

28. See Lind­say v. Wrecked & Aban­doned Ves­sel R.M.S. Titan­ic, 1999 U.S. Dist. LEXIS 15837, at *15–18 (S.D.N.Y. Oct. 13, 1999) (hold­ing the direc­tor was the author of the film due to the lev­el of con­trol he had over the entire process to ensure his orig­i­nal ideas were being pro­duced, regard­less of the fact that he did not actu­al­ly film every part of the movie); Brod v. Gen. Pub­l’g Group, Inc., 32 Fed. App’x 231, 234 (9th Cir. 2002) (find­ing the author’s con­tri­bu­tion to the pho­tos and the cre­ation of the book were suf­fi­cient­ly orig­i­nal “even though he did not trig­ger the shutter”).

29. See Torah Soft Ltd., 136 F. Supp. 2d at 283 (soft­ware pro­gram that com­piles infor­ma­tion); Rear­den LLC, 293 F. Supp. 3d at 963 (soft­ware pro­gram that con­verts human expres­sion into CGI); Dig­i­tal Drilling Data Sys. LLC, No. 4:15-CV-02172, 2018 U.S. Dist. LEXIS 83158, at *20–21 (soft­ware pro­gram that ben­e­fits the oil industry).

30. For instance, in Torah Soft, the user’s input of a sin­gle word into the com­put­er soft­ware pro­gram was deter­mined to be mere­ly mar­gin­al, since it was the sys­tem that com­piled all the infor­ma­tion. See Torah Soft Ltd., 136 F. Supp. 2d at 283. In con­trast, when the cre­ative choic­es in the input con­tributed by the user was large­ly reflect­ed in the out­put, the court in Rear­den LLC v. Walt Dis­ney Co., held that con­tri­bu­tion by the user was not mere­ly mar­gin­al. See Rear­den LLC, 293 F. Supp. 3d at 963.

31. See supra note 28.

32. See Lind­say, 1999 U.S. Dist. LEXIS 15837, at *15–18.

33. See supra note 21.

34. See Notes of Com­mit­tee on the Judi­cia­ry, H.R. Rep. No. 1476, 94th Cong., 2d Sess. 51 (1976), reprint­ed in 1976 U.S.C.C.A.N. 5659, 5664 (“The his­to­ry of copy­right law has been one of grad­ual expan­sion in the types of works accord­ed pro­tec­tion.… Authors are con­tin­u­al­ly find­ing new ways of express­ing them­selves, but it is impos­si­ble to fore­see the forms that these new expres­sive meth­ods will take.”).

35. See, e.g., Gar­cia v. Google, Inc., 786 F.3d 733, 741 (9th Cir. 2015).

36. Melville B. Nim­mer & David Nim­mer, 1 Nim­mer on Copy­right § 2.03 (Matthew Ben­der, rev. ed. 2019) (sug­gest­ing that Con­gress made a delib­er­ate choice to leave the term author­ship undefined).

37. Id; see Gar­cia, 786 F.3d at 741.

38. See Bur­row-Giles Lith­o­graph­ic Co. v. Sarony, 111 U.S. 53, 56–60 (1884) (grant­i­ng copy­right pro­tec­tion for pho­tographs regard­less of the tech­no­log­i­cal and nov­el nature of cam­eras and photography).

39. Id. at 58.

40. See Dis­cov­er the Sto­ry Behind the Next Rem­brandt, The Next Rem­brandt, https://www.nextrembrandt.com/ (last vis­it­ed Feb 8, 2021) (describ­ing the process of cre­at­ing the Next Rem­brandt as entail­ing the cre­ation of a data­base of all of Rembrandt’s orig­i­nal works, which was then inputted into the AI program).

41. Id.

42. Id.

43. See Feist Publ’ns, Inc., 499 U.S. at 348 (deter­min­ing copy­right pro­tec­tion can be grant­ed in sit­u­a­tions where the cre­ator mere­ly selects, com­piles, or arranges the var­i­ous elements).

44. See Kel­ley v. Chi. Park Dist., 635 F.3d 290, 302 (7th Cir. 2011) (deny­ing copy­right pro­tec­tion for a park exhib­it cre­at­ed by a human).

45. Id.

46. See Torah Soft Ltd., 136 F. Supp. 2d at 283 (uti­liz­ing the lion’s share bal­anc­ing test to deter­mine whether the pro­gram user should gain copy­right pro­tec­tion for con­tribut­ing a mar­gin­al amount of work); Rear­den LLC, 293 F. Supp. 3d 963 (apply­ing the bal­anc­ing test to deter­mine pro­gram cre­ator could claim copy­right pro­tec­tion over the out­put); see also Dig­i­tal Drilling Data Sys. LLC , No. 4:15-CV-02172, at *20–21 (employ­ing the lion’s share test when deter­min­ing copy­right pro­tec­tion over the schema cre­at­ed by a soft­ware program).

47. See Rear­den, 293 F. Supp. 3d at 963.

48. Id.

49. Id.

50. See supra note 28.

51. Lind­say, 1999 U.S. Dist. LEXIS 15837, at *15–18.

52. Id.

53. See Cmty. for Cre­ative Non-Vio­lence v. Reid, 490 U.S. 730,737 (“As a gen­er­al rule, the author is the par­ty who actu­al­ly cre­ates the work, that is, the per­son who trans­lates an idea into a fixed, tan­gi­ble expres­sion enti­tled to copy­right protection.”).

54. See Bur­row-Giles Lith­o­graph­ic Co., 111 U.S. at 56–60 (grant­i­ng pro­tec­tion for photographs).

55. See U.S. Con­st. art. I, § 8, cl. 8. (“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”).

56. See, e.g., Sony Corp. of Am. v. Uni­ver­sal City Stu­dios, Inc., 464 U.S. 417, 429 (1984); Weiss­mann v. Free­man, 868 F.2d 1313, 1325 (2d Cir. 1989) (find­ing the incen­tive inter­ests an impor­tant fac­tor in deter­min­ing copy­right pro­tec­tion due to its effect of caus­ing con­tin­ued research); see also Maz­er v. Stein, 347 U.S. 201, 219 (1954) (hold­ing copy­right pro­tec­tion for stat­uettes used as the base of a lamp in part due to the incen­tive to fur­ther research, dis­sem­i­nate ideas, and cre­ate art).

57. Twen­ti­eth Cen­tu­ry Music Corp. v. Aiken, 422 U.S. 151, 156 (1975) (dis­cussing the advent of the com­mer­cial radio and the grant­i­ng of copy­right infringe­ment to the broad­cast of copy­right­ed musi­cal compositions).

58. See Daryl Lim, AI & IP Inno­va­tion & Cre­ativ­i­ty in an Age of Accel­er­at­ed Change, 52 Akron L. Rev. 813, 815 (2018)(“Excluding AI-gen­er­at­ed work for pro­tec­tion leaves an eco­nom­ic lacu­na, which U.S. IP pol­i­cy is loath to do. Allow­ing AI-gen­er­at­ed works to fall into the pub­lic domain reduces the incen­tive to invest in the growth of the industry.”).

59. E.g., Gar­cia v. Google, Inc., 786 F.3d 733, 741 (9th Cir. 2015) (cit­ing Melville B. Nim­mer & David Nim­mer, 1 Nim­mer on Copy­right § 2.03 (Matthew Ben­der, rev. ed. 2019)).

60. See Feist Publ’ns, Inc. v. Rur­al Tel. Serv. Co., 499 U.S. 340, 111 (1991) (set­ting out the require­ments for copy­right protection).

61. See supra note 6.

62. See supra notes 57–59.