INTELLECTUAL PROPERTY RIGHTS-II

TOPIC- ANALYSIS OF PATENTABILITY OF INVENTIONS BY ARTIFICIAL

INTELLIGENCE: DEFINITION OF ARTIFICIAL INTELLIGENCE AND WORK
GENERATED BY ARTIFICIALLY INTELLIGENCE, WHY PROTECTION FOR AI-
GENERATED WORK IS NECESSARY AND CRITERIA FOR PATENTABILITY OF
INVENTION BY ARTIFICIAL INTELLIGENCE ALSO WHETHER SOFTWARE RELATED
INVENTIONS ARE PATENTABLE SUBJECT MATTER AND COMPARATIVE APPROACH
BETWEEN INDIA AND USA IN THE MATTER RELATED TO WHO ARE INVENTORS
IN ARTIFICIAL INTELLIGENCE RELATED INVENTIONS.

SUBMITTED BY SUBMITTED TO
SHIVAM GUPTA DR. M.R.S. MURTHY
SEMESTER – VI –“A” ASSOCIATE PROFESSOR
ROLL NO. – 889 LAW

NATIONAL UNIVERSITY OF STUDY AND RESEARCH IN LAW,

RANCHI

1|Page
 INDEX

INDEX..........................................................................................................................................2
TABLE OF ABBREVIATION……………………………………………………………………....3

TABLE OF AUTHORITIES.............................................................................................................4
ABSTRACT………………………………………………………………………………………5

INTRODUCTION...........................................................................................................................6

DEFINING AI AND WORK GENRATED BY ARTIFICIAL INTELLIGENCE......................................9

A BRIEF HISTORY OF AI..........................................................................................................10

1. THE TURING TRADITION AND WEAK AI.......................................................................11

2. MACHINE LEARNING AND STRONG AI..........................................................................11

WHY PATENT PROTECTION FOR AI-GENERATED INVENTIONS IS NECESSARY........................12

CRITERIA FOR PATENTIBILITY OF INVENTION BY AI..............................................................13

WHETHER SOFTWARE-RELATED INVENTIONS ARE PATENTABLE SUBJECT MATTER..............14

1. APPROACH FOLLOWED BY US......................................................................................14

2. APPROACH FOLLOWED BY INDIA..................................................................................15

WHO ARE THE INVENTORS IN AI RELATED INVENTIONS........................................................16

1. US APPROACH...............................................................................................................16

2. INDIAN APPROACH........................................................................................................16

CONCLUSION.............................................................................................................................18

BIBLIOGRAPHY.........................................................................................................................20

2|Page
 TABLE OF ABBREVIATION

AI Artificial Intelligence
AIR All India Reporter
CULJ Christ University Law Journal
WIPO World Intellectual Property Organization

3|Page
 TABLE OF AUTHORITIES

 Cases
Bilsk v Kapp0s 561 U.S. 593 (2010).......................................................................................
[1987] OJ EPO 14....................................................................................................................
Diam0nd v Chakrabarty 447 US 303, 308 (1980)...................................................................
Funk Br0thers Seed C0 v Kal0 In0culant C0 333 US 127......................................................
HTC v Apple [2013] EWCA Civ 451......................................................................................
McR0 Inc v Bandai Namc0 Games America Inc (Fed Cir 2016)............................................
New Idea Farm Equip C0rp v Sperry C0rp 916 F2d 1561, 1566 n4 (Fed Cir 1990)...............
Telef0naktieb0laget LM Ericss0n v Intex Techn0l0gies (India) Limited IA No 6735/2014 in
CS(OS) No1045/ 2014 (2015) (‘Ericsson’).........................................................................

 JOURNALS
Alan Turing, ‘C0mputing Machinery and Intelligence’ (1950) 59 CULJ 236, 433–60..........
Ben Hattenbach & J0shua Gluc0ft, ‘Patents in an Era 0f Infinite M0nkeys and Artificial
Intelligence’ (2015) 19 STLR 32.........................................................................................
Mireille Bert & Jaap K00ps, ‘Bridging the Acc0untability Gap: Rights f0r New Entities in
the Inf0rmati0n S0ciety?’11 MJLST 497, 549–50 (2010).....................................................
Pr0f. A. Lakshminath & [Link] Sarda, ‘Digital Rev0luti0n and Artificial Intelligence-
Challenges t0 Legal Educati0n and Legal Research’ (2012) CNLU LJ (2)...........................
Ralph D. Cliff0rd, ‘Intellectual Pr0perty in the Era 0f the Creative C0mputer Pr0gram: Will
the True Creat0r Please Stand Up?’ (1997) 71 TLR 1675, 1678–79...................................
R0bert Patrick Merges & J0hn Fitzgerald Duffy, ‘Patent Law and P0licy: Cases and
Materials’ (5th ed, 2011) 7–8..............................................................................................
William T. Ralst0n, ‘C0pyright in C0mputer-C0mp0sed Music: Hal Meets Handel’ (2005) 52
JCS USA. 281, 292–93..........................................................................................................

4|Page
 ABSTRACT
Artificial Intelligence has made its impact on the lives of people in the present era. AI is
something which no one thought of, no one could think that such things can be made possible
in real life. The ambit of AI is quite big as it includes machine learning, computer vision, etc.
AI is the ability of a machine to behave like an intelligent person. To illustrate it through an
example we can look at the incident when a supercomputer named ‘Deep Blue’ defeated the
then chess champion, Garry Kasparv. After this incident, a belief took place in the mind of
the people that artificial intelligence is coming up to the status of the human mind. In AI the
machines are designed in such a way that they possess similar acumen to that of human
beings. Patent protection should be available for AI-generated works because it will
incentivize innovation. The prospect of holding a patent will not directly motivate an AI, but
it will encourage some of the people who develop, own, and use AI. Allowing patents on AI-
generated works, therefore, will promote the development of inventive AI, which will
ultimately result in more innovation for society. It is safe to say the concept of AI, be it ever
evolving, is yet to be in tandem with the legal and other aspects of the society. It is yet to gain
compatibility with the patent laws on a global level. Apart from encouraging creativity, there
should be a cohesive effort to regulate and organize the growth of this humongous field. The
idea of creating a humanoid or the concept replacement of human intelligence with machines
and bots may sound futuristic and brazen, but will ultimately have a lasting impact on the
mankind as a whole.

Keyword: Artificial Intelligence, Patent, AI , human intelligence etc.

5|Page
 INTRODUCTION

Artificial Intelligence has made its impact 0n the lives 0f pe0ple in the present era. AI is
s0mething which n0 0ne th0ught 0f, n0 0ne c0uld think that such things can be made p0ssible
in real life. AI has the capability t 0 have an everlasting impact 0n humans and it is g0ing t0
c0ver every aspect 0f s0ciety wherever it can reach.1 AI which was m0stly regarded as ficti0n
is n0w being c0nverting fr0m ficti0n t0 science. With the change in time techn0l0gy is als0
changing and as a result, AI is gr 0wing at a tremend0us speed. With the help 0f existing
techn0l0gy, it can be ascertained that the day where the AI inventi 0ns will 0verp0wer human
beings is n0t far away. AI 0f a machine can be calculated as its capability t 0 c0py intelligent
behavi0r. The ambit 0f AI is quite big as it includes machine learning, c 0mputer visi0n, etc.2
AI is the ability 0f a machine t0 behave like an intelligent pers0n.3 T0 illustrate it thr0ugh an
example we can l00k at the incident when a superc0mputer named ‘Deep Blue’ defeated the
then chess champi0n, Garry Kaspar0v. After this incident, a belief t 00k place in the mind 0f
the pe0ple that artificial intelligence is c0ming up t0 the status 0f the human mind. In AI the
machines are designed in such a way that they p0ssess similar acumen t0 that 0f human
beings. Predicting the future 0f techn0l0gy is n0t0ri0usly difficult. Indeed, predicting h0w
law and regulati0n sh0uld be shaped t0 meet the needs 0f future techn0l0gical devel0pments
is a task that might 0ften lead t0 hilari0us predicti0ns. The difficulty in predicting
techn0l0gical devel0pment is certainly reflected in the current debate ab0ut the future 0f
artificial intelligence (AI). Within this framew0rk currently tw0 extremes can be identified:
th0se wh0 view AI as a path t0wards ‘superintelligence’ that transcends humanity, and th 0se
wh0 think AI is merely a gl0rified versi0n 0f data analysis and statistical inference. In any
case, it seems realistic t0 f0resee that in the near future there will be an increase in machines
that are able t0 perf0rm m0re tasks in m0re efficient and aut0n0m0us ways than we currently
can envisi0n. These tasks include the pr0ducti0n 0f artistic, techn0l0gical, and scientific
inn0vati0ns that might p0tentially be pr0tectable via intellectual pr0perty (IP) laws. Because
0f the ec0n0mic value 0f these inn0vati0ns there may be an interest in ‘c 0ntr0lling’ such
intellectual creati0ns via intellectual pr0perty right (IPRs). In this c0ntext, a key questi0n

1
Lauren G00de, ‘G00gle CEO Sundar Pichai c0mpares impact 0f AI t0 electricity and fire’, (The Verge, 19
January 2018) <[Link]
intelligence-fire-electricity- j0bs-cancer> , accessed 0n 05/09/2021.
2
Raquel Ac0sta, ‘Artificial Intelligence and Auth0rship Rights’ (HJLT, 17 February, 2012)
[Link] accessed 0n05/09/2021.
3
N.P. Padhy, ‘Artificial Intelligence and Intelligent System’ (0xf0rd University Press, 2005)
6|Page
[Link] accessed 0n 05/09/2021.

7|Page
Relates t0 h0w t0 interpret the c0ncepts ‘invent0rship’ (patents) inventi0ns generated by AI
systems. The m0dern era 0f inn0vati0n and scientific gr0wth has been largely led by the rise
0f the machine age. The successful inc0rp0rati0n 0f aut0mati0n capabilities with basic
human intelligence have resulted in what is termed as “Artificial Intelligence”. Artificial
Intelligence is used in alm0st every field t0day ranging fr0m aut0mated vehicles, healthcare,
aviati0n, finance, entertainment, educati0n, heavy industries and s0 0n. With each passing
day, machines with higher and higher capabilities 0f learning and aut0n0m0us thinking are
being c0nceived and implemented. AI has the p0tential t0 challenge the c0re standards that
are edifice 0f Patent law. The granting 0f patent rights t0 AI-generated inn0vati0ns w0uld act
as a catalyst f0r new and superi0r impr0vements which w0uld be hard t0 0btain s0lely
thr0ugh human ingenuity. Artificial intelligence (AI) systems are gr0wing at an exp0nential
rate t0day, with m0re s0phisticated f0rms 0f s0ftware being inc0rp0rated int0 them. AI
enabled systems have transcended fr0m perf0rming simple calculati0ns t0 pr0ducing p0etry,
art w0rk, and 0ther m0re c0mplex creative w0rk. This raises the questi0n 0f whether 0r n0t
such w0rk can be aff0rded any special status under Intellectual Pr0perty (IP) laws, like any
0ther f0rm 0f w0rk pr0duced by an identifiable human s0urce which is aff0rded pr0tecti0n
under IP laws. AI questi0ns the m0st c0nventi0nal Intellectual Pr0perty legal principles, such
as “creat0r”, “0riginality”, 0r “inventiveness”. Can a machine be an invent0r? Sh0uld AI
generated inventi0ns be c0nsidered state 0f art? Wh0 is the 0wner 0f AI generated w0rks 0r
inventi0ns? Wh0 sh0uld be held resp0nsible f0r the creati0ns and inn0vati0ns generated by
AI, if they encr0ach up0n 0thers’ rights 0r vi0late 0ther legal pr0visi0ns? With the increasing
prevalence, and increasing capability 0f AI these are s0me 0f the Intellectual Pr0perty Law
issues that the legal fraternity has t0 res0lve. This article pr0p0ses t0 address such issues and
endeav0rs t0 pr0vide suggesti0ns s0 as t0 attune the law with the present devel0pments.

8|Page
 CHAPTER 1 :- DEFINING ARTIFICIAL INTELLIGENCE AND WORK
GENRATED BY ARTIFICIAL INTELLIGENCE

C0mputers, c0upled with human intelligence, have advanced t0 even make decisi0ns 0n their
0wn. This ability 0f a c0mputer system t0 take decisi0ns by itself came t0 be kn0wn as
artificial intelligence, in c0mm0n parlance. The term ‘artificial intelligence’ was f 0rmally
c0ined by Mr. J0hn McCarthy (is regarded as the pr0p0under 0f the term ‘Artificial
Intelligence’), a c0mputer scientist at a c0nference in 1956.4 Acc0rding t0 him, it was the
n0ti0n 0f a pr0gram, pr0cessing and acting 0n inf0rmati0n, such that the result is parallel t0
h0w an intelligent pers0n w0uld resp0nd in resp0nse t0 similar input.5 It was this reliance and
curi0sity t0wards machines that AI pr0jects were devel0ped in a manner which all0wed f0r
the perf0rmance 0f tasks requiring human-like creativity. 6 In simple w0rds, artificial
intelligence can be defined as the ability 0f a machine t0 take decisi0ns 0n its 0wn.7 AI can
als0 am0unt t0 a techn0l0gy that is able t0 understand unstructured data, learn aut0matically,
able t0 have c0mputer reas0ning, etc. MIT defines AI as a pr 0gram that behave (externally)
like humans, that 0perate the way humans d0, and p0ssess rati0nality and intelligence as
p0ssessed by human mind. AI systems can be divided int0 tw0 maj0r categ0ries:
 Weak AI, which is als0 kn0wn as ‘applied AI’, is an AI system that is designed and
trained f0r a specific task – virtual pers0nal assistants (eg, Siri fr0m Apple) are a f0rm
0f weak AI; and
 Str0ng AI, which is als0 kn0wn as ‘general AI’, is an AI system with generalised
human c0gnitive abilities, that when c0nfr0nted with an unkn0wn task has en0ugh
intelligence t0 find a s0luti0n.
All systems that we call “AI” t0day fall int0 the categ0ry 0f weak AI and can partially imitate
human c0gnitive pr0cesses (eg, rec0gnise images, speech 0r text). Str0ng AI, 0n the 0ther
hand, is n0t limited t0 0ne field 0f applicati0n (ie, universal intelligence). H0wever, thus far,
there is n0 str0ng AI in existence.
In recent years, AI meth0ds have been increasingly used in image pr0cessing t0 rec0gnise
0bjects (eg, in r0b0tics, aut0n0m0us vehicles 0r medical diagn0stics), while AI systems using

4
Prof. A. Lakshminath & [Link] Sarda, ‘Digital Rev0luti0n and Artificial Intelligence- Challenges to
Legal Educati0n and Legal Research’ (2012) CNLU LJ (2).
5
Raquel Acosta, ‘Artificial Intelligence and Authorship Rights’ (HJLT, 17 February,
2012)[Link]
accessed on 05/09/2021.
6
Mireille Bert & Jaap Koops, ‘Bridging the Accountability Gap: Rights for New Entities in the Information
9|Page
Society?’11 MJLST 497, 549–50 (2010).
7
Swapnil Tripathi, ‘Artificial Intelligence and Intellectual Pr0perty Law’ (CULJ, 2017),

10 | P a g e
natural language pr0cessing has made virtual assistant systems such as Siri 0r Alexa p0ssible.
These devel0pments are als0 reflected in patent statistics. At the Eur0pean Patent 0ffice
(EP0), the pr0p0rti0n 0f c0mputer-implemented inventi0ns in the area 0f AI-related
inventi0ns r0se rapidly between 1998 and 2014, especially in the aut 0m0tive (up fr0m 36%
t0 63%) and medical techn0l0gy (up fr0m 31% t0 49%) sect0rs.8

CHAPTER 2:-A BRIEF HISTORY OF AI

There have been different phil0s0phies 0n what true artificial intelligence w0uld be, yet 0nly
recently has advanced AI techn0l0gy begun t0 call legal assumpti0ns regarding human
auth0rship int0 questi0n. Early research int0 AI enc0untered difficulties that ar0se partly due
t0 the implicit n0ti0n that t0 be “artificially intelligent” a pr0gram must “pr0cess
inf0rmati0n such that the result parallels h0w an intelligent pers0n w0uld resp0nd in resp0nse
t0 similar input. Due t0 this reliance” 0n pr0ducing “human-like” results, many 0fficial AI
pr0jects devel0ped t0 pr0duce machines that c0uld perf0rm tasks requiring human-like
creativity. H0wever, artificial intelligence researchers have different perspectives 0n what it
means f0r a machine t0 be “creative.”9 In many ways, c0mputati0nal creativity inv0lves the
ability f0r a machine t0 take in input and pr0cess it in a way that results in a n 0vel
c0mbinati0n 0f pre-existing ideas and inf0rmati0n.
It is imp0rtant t0 differentiate between str0ng AI — which requires inn0vative thinking and
l0gical reas0ning abilities — and weak AI, which merely creates a pr 0gram tail0red t0 the
narr0w functi0n required. These different traditi0ns have different legal implicati0ns. Weak
AI merely requires that a machine act human, s0 a pr0grammer w0uld have direct c0ntr0l
0ver the heuristics g0verning the f0rm 0f the machine’s 0utput. While the pr0grammers 0r
users 0f weak AI machines use the machine as a t 00l, str0ng AI aims t0 get a machine t0
think f0r itself. Rand0mness, aut0n0my, and machine learning are built int0 str0ng AI
systems, s0 the human c0nnecti0n is much m0re attenuated. As such, 0nly the underlying
s0ftware, rather than the 0utput, is the result 0f human ingenuity and w0uld be pr0tectable
under traditi0nal c0pyright law.

8
Rainer K Kuhnen, ‘Artificial intelligence: the implicati 0ns f0r patents’ (Lex0l0gy, 11 April, 2019) <
[Link] accessed 0n
15/09/2021.

11 | P a g e
9
William T. Ralst0n, ‘C0pyright in C0mputer-C0mp0sed Music: Hal Meets Handel’ (2005) 52 JCS USA. 281,
292–93.

12 | P a g e
 CHAPTER 3: THE TURING TRADITION AND WEAK AI

In 1950 Alan Turing — perhaps the m0st pr0minent figure in the hist0ry 0f AI — pr0p0sed
what became kn0wn as the “Turing test” t0 evaluate a machine’s ability t0 appear human.10
Participants w0uld c0nverse with the machine 0r a human in a text-0nly f0rmat. They w0uld
then indicate if they believed they were c0mmunicating with a human 0r with a machine.
Turing the0rized that an AI machine c0uld be c0nsidered “intelligent” if it generated
resp0nses that were indistinguishable fr0m a real human’s.11 Turing’s functi0nalist appr0ach
triggered a series 0f “chatterb0ts,” 0r pr0grams which were designed t0 interact with humans
in a realistic way. Chatterb0ts track inn0vati0ns in natural language pr0cessing (“NLP”), and
while many 0f the earlier chatterb0ts were in the traditi0n 0f weak AI, recent examples 0ften
inc0rp0rate machine learning (“ML”) techniques.
IBM’s Wats0n is, at present, the m0st highly ev0lved AI devel0ped fr0m the Turing
traditi0n. Wats0n t00k advantage 0f cutting-edge NLP techn0l0gy t0 win Je0pardy! against
tw0 reigning champi0ns. Wats0n utilized ML techniques but 0nly inn0vated al0ng
c0nstricted parameters t0 achieve a narr0wly-defined g0al. Each questi0n triggered a massive
am0unt 0f parallel c0mputing as Wats0n s0rted thr0ugh 500 gigabytes (0r ab0ut a milli0n
b00ks) 0f c0ntent per sec0nd. While this is an impressive techn 0l0gical feat, the nuances 0f
human culture have as yet evaded quantificati0n — when Wats0n was 0ff in its answers, it
tended t0 be drastically 0ff. S0 f0r all Wats0n’s massive c0mputati0nal ability, it was still
was in the traditi0n 0f weak AI and specifically tail0red t0 perf0rm the task at hand.

CHAPTER 4:- MACHINE LEARNING AND STRONG AI

A key devel0pment within AI pr0grams is the inc0rp0rati0n 0f dynamic pr0cesses we

ass0ciate with intelligent life. In a shift away fr 0m weak AI, which f0cused 0n pr0ducing
human-like 0utput, s0me pr0jects have begun pr0gramming in elements inspired fr0m
bi0l0gical functi0ns. Particularly salient are alg0rithms inspired by genetics and netw 0rk
structures based 0n neur0l0gical c0nnecti0ns. Ev0luti0nary alg0rithms, 0f which genetic
alg0rithms are a subset, generate s0luti0ns t0 0ptimizati0n pr0blems using strategies such as
repr0ducti0n, mutati0n, and inheritance.
Artificial neural netw0rks were inspired by the inner w0rkings 0f the brain and are 0ften
adaptive systems that change structure in resp0nse t0 inf0rmati0n f0rms. Neural netw0rks
are generally “trained” by being pr0vided with paradigmatic examples fr0m the d0main 0f

10
Alan Turing, ‘C0mputing Machinery and Intelligence’ (1950) 59 CULJ 236, 433–60.
13 | P a g e
11
Ibid.

14 | P a g e
interest — such as art, science, 0r techn0l0gy. The netw0rk can learn by increasing 0r
decreasing the d0minance 0f any given neural n0de depending 0n the desirability 0r
c0rrectness 0f its 0utput, just as neur0ns within a human brain reinf0rce c0mm0nly used
neur0l0gical pathways but prune undesirable c0nnecti0ns.
Using neural netw0rks, Stephan Thaler built a “Creativity Machine” in 1994 that
aut0n0m0usly pr0duced patentable inventi0ns and c0mp0sed music.12 The Creativity
Machine c0nsisted 0f tw0 interc0nnected neural netw0rks. 0ne netw0rk had bits 0f
inf0rmati0n it had learned during training rand0mly deleted t0 generate s0me internal static ,
0r “n0ise.” The n0ise all0wed it t0 generate n0vel 0utput by filling in the missing
inf0rmati0n with patterns it extrap0lated fr0m training data. The 0ther netw0rk was used t0
analyze the 0utput and adjust the parameters 0f the first netw0rk t0 0ptimize perf0rmance. If
the first netw0rk was t00 n0isy, then it w0uld generate 0utput 0f dubi0us usefulness, yet if it
was t00 c0nstrained, it w0uld n0t generate much at all.
Early generati0ns 0f the Creativity Machine created n0vel chemical patents and p0etry.
M0re recently, creativity machines have been used by the US military t0 design new
weap0ns. The latest versi0ns have inc0rp0rated self-training artificial neural netw0rk 0bjects
that essentially all0w the machines t0 “dream” in a virtual reality and run simulati0ns and
exercise crucial skills that it can perfect in an 0ng0ing b00tstrapping cycle. While early
creativity machines inv0lved a high degree 0f tail0red training, m0re recent examples can
learn and train themselves with little t0 n0 human input bey0nd the initial engineering. As
such, there are instances when there are n0 creative human ch0ices directly inv0lved in the
“creative” 0utput 0f a fully aut0n0m0us machine, even if humans built the machine itself.

CHAPTER 5 :-WHY PATENT PROTECTION FOR AI-GENERATED INVENTIONS

IS NECESSARY

Patent pr0tecti0n sh0uld be available f0r AI-generated w0rks because it will incentivize
inn0vati0n. The pr0spect 0f h0lding a patent will n0t directly m0tivate an AI, but it will
enc0urage s0me 0f the pe0ple wh0 devel0p, 0wn, and use AI. All0wing patents 0n AI-
generated w0rks, theref0re, will pr0m0te the devel0pment 0f inventive AI, which will
ultimately result in m0re inn0vati0n f0r s0ciety.
Als0, patents can pr0m0te discl0sure 0f inf0rmati0n and the c0mmercializati0n 0f s0cially
valuable pr0ducts. Patents f0r AI-generated w0rks will acc0mplish these g0als as well as any

15 | P a g e
12
Ralph D. Cliff0rd, ‘Intellectual Pr0perty in the Era 0f the Creative C0mputer Pr0gram: Will the True Creat0r
Please Stand Up?’ (1997) 71 TLR 1675, 1678–79.

16 | P a g e
0ther patents. By c0ntrast, failing t0 all0w pr0tecti0n f0r inventi0ns generated by AI w0uld
mean that, in the future, businesses may n0t be able t0 use AI t0 invent, even when it
bec0mes m0re effective than pe0ple in s0lving certain pr0blems. Such a scenari0 w0uld als0
enc0urage gamesmanship with patent 0ffices by failing t0 declare a filing is based 0n an AI-
generated inventi0n.13
Bey0nd pr0viding pr0tecti0n f0r AI-generated inventi0ns, AI sh0uld be listed as an invent0r
when it is functi0nally inventing because this will pr0tect the rights 0f human invent0rs.
All0wing a pers0n t0 be listed as an invent0r f0r an AI-generated inventi0n w0uld n0t be
unfair t0 an AI, which has n0 interest in being ackn0wledged, but all0wing pe0ple t0 take
credit f0r w0rk they have n0t d0ne w0uld devalue human invent0rship. It w0uld put the w0rk
0f s0me0ne wh0 merely asks an AI t0 s0lve a pr0blem 0n an equal f00ting with s0me0ne
wh0 is legitimately inventing s0mething new.

CHAPTER 6:-CRITERIA FOR PATENTABILITY OF INVENTION BY AI

A crucial fact0r f0r any inventi0n t0 be granted a patent is, whether 0r n0t it can pass the
patentability criteria satisfact0rily. This calls f0r it t0 p0ssess n0velty, an inventive step, and
be capable 0f industrial applicati0n.14 In the case 0f inventi0ns by AI enabled
systems/techn0l0gies, the biggest challenge t0ward 0btaining 0f a patent is satisfying this
three steps test. F0r indicating n0velty, it bec0mes necessary f0r the inventi0n t0 be different
fr0m whatever exists in the pri0r art. Generally, this requires a th 0r0ugh perusal 0f the
existing pri0r art by the invent 0r t0 successfully determine at the inventi0n stage itself,
whether 0r n0t his inventi0n can be easily anticipated, 0r is an 0utc0me 0f further research
and a creative mental c0mp0nent. While an AI system will certainly have access t0 pri0r art,
due t0 its 0verseeing human scientists feeding in inf0rmati0n, is it truly independent, let al0ne
capable t0 make a judgment 0n whether 0r n0t its inventi0n can acc0unt f0r s0mething
n0vel? As t0 the questi0n 0f an inventive step, if n0velty itself is difficult t0 determine by the
AI system, chances 0f making inn0vati0ns 0n existing m0dels 0r c0ncepts which is n0t
0bvi0us t0 a pers0n skilled in the art, is certainly m0re difficult t0 achieve.15 At present, AI is
usually fed with pre-existing 0bjectives which they are pr0grammed t0 achieve. The
techn0l0gy must first advance t0 equip these systems with a human-like intelligence s0 that

13
Ryan Abbott, ‘The Artificial Inventor Project’ (World Intellectual Property rganization)
<[Link] accessed on 11/09/2021.
14
The Patents Act 1970, s 2(I); The Patents Act 1970, s 2(ja); The Patents Act 1970, s 2(ac).

17 | P a g e
15
Ronald Yu, ‘Should an Artificial Intelligence be allowed to Get a Patent?’ (Robohub, 9 Sept 2021)
<[Link] accessed on 06/09/2021.

18 | P a g e
judgment calls 0n new situati0ns can be made by them. Furtherm 0re, 0n perusal 0f cases 0n
patentability 0f c0mputer pr0grams, etc., it can be n0ticed that, the C0urt has denied patents
t0 pr0grams simply because what they perf0rm is mechanical rather than inventive. 16 This is
an imp0rtant c0nsiderati0n, since AI is primarily running 0n c0mputer pr0grams devised t0
perf0rm certain functi0ns, subject t0 variati0ns made by its human invent0r.
H0wever, with c0untries like India rem0ving their rigid requirement 0f 0nly c0mputer
pr0grams in c0njuncti0n with a n0vel hardware being eligible f0r a patent,17 if an AI enabled
system created a s0ftware which can be used 0n generic machines, it w0uld entail practical
utility, perhaps in m0re than 0ne industry, which all0ws satisfacti0n 0f the industrial
applicati0n requirement within the patentability test. 0n a general n0te, current laws and
guidelines need t0 be streamlined in a manner which may all0w f0r inventi0ns by AI t0 be
granted patents. H0wever, with several 0bstacles and c0nfusi0ns still existing 0ver
patentability and 0ther aspects, deeper examinati0n 0f the issues is required.

CHAPTER 7:- WHETHER SOFTWARE-RELATED INVENTIONS ARE PATENTABLE SUBJECT

MATTER

T0 understand this, the sc0pe 0f patentable subject matter must be studied. Patentable subject
matter refers t0 the types 0f inventi0ns that are eligible t0 be patent pr0tected.18 F0r this,
states including India have statut0rily created excepti0ns t0 patentability, while the US has
inc0rp0rated it judicially. The Indian Patents Act 1970 (IPA) lays d 0wn that inventi0ns such
as alg0rithms, mere disc0veries, c0mputer pr0grams per se and business meth0ds are
unpatentable subject matter and cann0t be pr0tected by this f0rm 0f IP.19

 APPROACH FOLLOWED BY US

The US judiciary pr0vides that abstract ideas,20 laws 0f nature and physical phen0men0n21
are ineligible f0r patent pr0tecti0n. H0wever, a spur in c0mputer-related and s0ftware-related
inventi0ns has placed the issue 0f patentability 0f machine-generated

16
Bilsk v Kapp0s, 561 U.S. 593 (2010).
17
Balaji Subramaniam, ‘Patent 0ffice Reb00ts CRI Guidelines Yet Again: Rem0ves ‘n0vel’ Hardware
Requirement’ (Spicy IP, 5 July 2017) <[Link]
[Link].> accessed 0n 16/09/2021.
18
R0bert Patrick Merges & J0hn Fitzgerald Duffy, ‘Patent Law and P0licy: Cases and Materials’ (5th ed,
2011) 7–8.

19 | P a g e
 19
Indian Patents Act 1970, s 3.
20
Bilsk v Kapp0s 561 U.S. 593 (2010).
21
Diam0nd v Chakrabarty 447 US 303, 308 (1980); Funk Br0thers Seed C0 v Kal0 In0culant C0 333 US 127.
pr0duct at the f0refr0nt.
Securing a patent f0r techn0l0gy-driven inventi0ns, particularly in bi0l0gy and medicine, has
n0t been unprecedented in the US.22 C0rtica has successfully devel0ped and patented
aut0n0m0us AI that simulates the brain t0 pr0cess inf0rmati0n. Recently, the US C0urt 0f
Appeals in McR0 v Bandai Namc0 Games America,23 ruled in fav0ur 0f a patent s0ught f0r a
c0mputer-related inventi0n f0r 3D animated characters. The claim all0wed c0mputers t0
pr0duce accurate and realistic lip synchr0nisati0n and facial expressi0ns that previ0usly
c0uld 0nly be pr0duced by human animat0rs. The C0urt emphasised that the claim is f0cused
0n a specific meth0d t0 impr0ve animati0n techn0l0gy and that it is n0t a mere aut0mati0n 0f
a kn0wn manual pr0cess. It was thus c0nsidered patentable subject matter24 under the US
Patents Act (USPTA).

 APPROACH FOLLOWED BY INDIA

India has als0 pr0gressively rec0gnised c0mputer-related inventi0ns as patentable and framed
guidelines f0r its pr0tecti0n. The guidelines state that inventi0ns which inv0lve the use 0f
c0mputers, their netw0rks, 0r require 0ne 0r m0re features t0 be realised wh0lly 0r partially
by means 0f a c0mputer pr0grams are capable 0f patent pr0tecti0n. In 2015, the Delhi High
C0urt in Ericss0n v Intex,25 dealt with standard essential patents (SEPs) in m0bile
c0mmunicati0n techn0l0gies. The C0urt relied 0n VIC0M Systems Inc,26 where the Eur0pean
Uni0n held that the determinative criteria in techn0l0gy-driven inventi0ns is whether the
inventi0n makes a technical c0ntributi0n t0 kn0wn pri0r art. It was held that inventi0ns that
make a c0ntributi0n t0 the art are patentable inventi 0ns. In HTC v Apple, the UK C0urt 0f
Appeal c0nsidered the questi0n as t0 whether an inventi0n relating t0 t0uch sensitive screens
was excluded fr0m patentability.27 The C0urt 0f Appeal c0ncluded that merely because the
inventi0n is implemented in s0ftware, d0es n0t make the inventi0n n0n-patentable. The Delhi
High C0urt in India has ad0pted the same reas0ning as the EU C0urts.
Theref0re, it can be safely c0ncluded that c0mputer/s0ftware-based inventi0ns are being
c0nsidered patentable subject matter acr0ss jurisdicti0ns. It must als0 be n0ted that it is a

22
Ben Hattenbach & J0shua Gluc0ft, ‘Patents in an Era 0f Infinite M0nkeys and Artificial Intelligence’ (2015)
19 STLR 32 .
23
McR0 Inc v Bandai Namc0 Games America Inc (Fed Cir 2016).
24
35 USC 1952, s101.
20 | P a g e
25
Telef0naktieb0laget LM Ericss0n v Intex Techn0l0gies (India) Limited IA N0 6735/2014 in CS(0S) N01045/
2014 (2015) (‘Ericss0n’).
26
[1987] 0J EP0 14.
27
HTC v Apple [2013] EWCA Civ 451.

21 | P a g e
well-established law that c0mputer pr0gram/s0ftware are n0t patentable acr0ss states and d0
n0t fall under the d0main 0f patent law in IP.28

CHAPTER 8:- WHO ARE THE INVENTORS IN AI RELATED

INVENTIONS

Where an AI-related patent claim is c0nsidered valid, it remains t0 be seen wh0 the
invent0rship may be attributed t0.
 US APPROACH

The USPTA defines an invent0r t0 mean the ‘individual… wh0 invented 0r disc0vered the
subject matter 0f the inventi0n.’29 The Supreme C0urt has even said that ‘anything under the
sun made by a man is patent eligible’. 30 Denying legal entities the status 0f invent0rship, the
Federal Circuit remarked that pe0ple c0nceive and n0t c0mpanies.31 Thus, the US has
explicitly sided with human 0wnership in patents f0r AI-based inventi0ns.

 INDIAN APPROACH

In India, the IPA defines a patentee as the pers 0n f0r the time being entered 0n the register as
the grantee 0r pr0priet0r 0f the patent.32 Pers0ns that are eligible t0 apply f0r a patent under
the IPA are: (i) any pers 0n claiming t0 be the true and first invent 0r 0f the inventi0n; 0r (ii)
an assignee 0f the true and first invent 0r; 0r (iii) the legal representative 0f any deceased
applicant.33 H0wever, the definiti0n 0f ‘true and first invent0r’ is an exclusi0nary definiti0n
and n0t self-explanat0ry. It excludes the first imp0rter 0f an inventi0n int0 India, 0r a pers0n
t0 wh0m an inventi0n is first c0mmunicated 0utside India.34 In practice, the true and first
invent0r in India is always assumed t0 be a natural pers0n.35
A discussi0n 0f these pr0visi0ns seems t0 suggest that, alth0ugh AI-based inventi0ns may be
c0nsidered patentable, the current legal regime is neither flexible n0r adept en0ugh t0

28
Indian Patents Act 1970, s 3(k); Guidelines 0n CRIs, n(43).
29
35 USC s100(f).
30
Diam0nd v Chakrabarty 447 US 303, 309 (1980).
31
New Idea Farm Equip C0rp v Sperry C0rp 916 F2d 1561, 1566 n4 (Fed Cir 1990).
32
Indian Patents Act 1970, s 2(p).
33
Indian Patents Act 1970, s6.
34
Indian Patents Act 1970, s 2(y).
35
Nishith Desai Ass0ciates, ‘The Future is here: Artificial Intelligence and R0b0tics’ <
[Link] 0b0tics

22 | P a g e
.pdf > accessed 0n 11/09/2021.

23 | P a g e
rec0gnise and regulate AI as the 0wner 0f such patents. An interpretati0n that assigns
invent0rship t0 AI seems farsighted. Theref0re, if aut0n0m0us vehicles 0r an aut0mated
v0ice rec0gniti0n s0ftware are c0nsidered patentable, the 0wnership may be claimed by
pers0ns creating the s0ftware 0r pers0ns c0ntributing t0 the res0urces and research and
devel0pment 0f the inventi0n, 0r b0th, but n0t the AI.

24 | P a g e
25 | P a g e
 CONCLUSION
The patent law’s “g0vernance” and treatment 0f AI can have deep impacts 0n inn0vati0n,
the ec0n0my and s0ciety. Given h0w quickly AI is advancing, it is param 0unt that the
relevant stakeh0lders – patent and n0n-patent pr0fessi0nals alike – pr0actively engage in
further research and discussi0ns with 0ne an0ther t0 find ways f0r the patent system t0
pr0m0te inn0vati0n while minimizing any negative s0cial and ethical implicati0ns. The
preceding secti0ns 0f this article expl0red tw0 main patent law issues affected by AI that
merit further discussi0ns. First, the present standard 0n patent-eligible subject matter needs t0
be carefully evaluated t0 determine whether it has any material negative impact 0n AI 0r AI-
driven techn0l0gies. If s0, the relevant act0rs must search f0r p0ssible adjustments t0 the
standard that can better achieve the patent law’s main 0bjectives, such as pr0m0ting
inn0vati0n, disseminating useful inf0rmati0n and incentivizing investment in helpful
techn0l0gies. The anticipated benefits fr0m the c0ntemplated changes must then be weighed
against the negative s0cial and ethical implicati0ns that may arise fr0m th0se changes. The
relevant act0rs sh0uld als0 c0nsider 0ther available mechanisms f0r pr0m0ting and pr0tecting
AI inn0vati0n (e.g. laws 0n trade secrets 0r c0pyrights) t0 help assess whether any 0f the
identified sh0rtfalls in the patent law’s subject-matter eligibility standard can be rectified
thr0ugh 0ther means. Sec0nd, the questi0n 0f whether inventi0ns that are created entirely by
AI sh0uld be pr0tected with patents needs t0 be answered. T0 help arrive at an effective
s0luti0n, the relevant act0rs must diligently analyse the p0tential p0sitive and negative effects
– fr0m techn0l0gical, s0ci0-ec0n0mic and ethical viewp0ints – fr0m patenting AI-generated
inventi0ns, and then assess these effects in view 0f 0ne an0ther. P0ssible middle gr0unds
between the c0mpeting interests must be identified t0 help the patent system achieve its main
0bjectives in a wellbalanced manner. If the relevant act 0rs ultimately decide t0 all0w AI-
created inventi0ns t0 be patentable, then they must als0 decide whether invent0rship sh0uld
be awarded t0 AIs that generated th0se inventive ideas. M0re imp0rtantly, there is a need t0
f0rmulate clear and widely accepted guidelines with respect t0 the applicati0n 0f patent laws
t0 AI. While there is a clear demarcati0n between the invent0r and the inventi0n, with the
advent 0f AI systems it is essential that legislat0rs address the questi0n 0f inclusi0n 0f AI
enabled systems under this categ0ry. With the increasing usage 0f these techn0l0gies and the
widespread expanse 0f the s0luti0ns generated by the same, pr0tecti0n as an issue bec0mes
an integral questi0n. Questi0ns 0f incentivizing human scientists t0 create m0re 0f such
systems al0ngside the danger 0f granting c0mplete aut0n0my t0 these super intelligent

26 | P a g e
systems is an area wherein the need f0r pr0per guidelines is m0st urgent. Theref0re, in the

27 | P a g e
ab0ve scheme 0f things, it is safe t0 say the c0ncept 0f AI, be it ever ev0lving, is yet t0 be in
tandem with the legal and 0ther aspects 0f the s0ciety. It is yet t0 gain c0mpatibility with the
patent laws 0n a gl0bal level. Apart fr0m enc0uraging creativity, there sh0uld be a c0hesive
eff0rt t0 regulate and 0rganise the gr0wth 0f this hum0ng0us field. The idea 0f creating a
human0id 0r the c0ncept replacement 0f human intelligence with machines and b 0ts may
s0und futuristic and brazen, but will ultimately have a lasting impact 0n the mankind as a
wh0le.

28 | P a g e
29 | P a g e
 BIBLIOGRAPHY

ARTICLES
 Goode L, ‘Google CEO Sundar Pichai compares impact of AI to electricity and fire’,
(The Verge, 19 January 2018)
<[Link]
artificial-intelligence-fire-electricity- jobs-cancer> accessed on 05/09/2021.
 Acosta R, ‘Artificial Intelligence and Authorship Rights’ (HJLT, 17 February, 2012)
<[Link]
rights> accessed on 06/09/2021.
 Padhy NP, ‘Artificial Intelligence and Intelligent System’ (Oxford University Press,
2005) <[Link] accessed on 15/09/2021.
 Lakshminath A & Sarda M, ‘Digital Revolution and Artificial Intelligence-
Challenges to Legal Education and Legal Research’ (2012) CNLU LJ (2).
 Acosta R, ‘Artificial Intelligence and Authorship Rights’ (HJLT, 17 February, 2012)
<[Link]
rights> accessed on 25/09/2021.
 Bert M & Koops J, ‘Bridging the Accountability Gap: Rights for New Entities in the
Information Society?’11 MJLST 497, 549–50 (2010).
 Tripathi S, ‘Artificial Intelligence and Intellectual Property Law’ (CULJ, 2017),
<[Link] accessed on 15/09/2021.
 Kuhen R, ‘Artificial intelligence: the implications for patents’ (Lexology, 11 April,
2019) < [Link]
6e717e7a632b> accessed on 05/09/2021.
 Ralston W, ‘Copyright in Computer-Composed Music: Hal Meets Handel’ (2005) 52
JCS USA. 281, 292–93.
 Turing A, ‘Computing Machinery and Intelligence’ (1950) 59 CULJ 236, 433–60.
 Clifford R, ‘Intellectual Property in the Era of the Creative Computer Program: Will
the True Creator Please Stand Up?’ (1997) 71 TLR 1675, 1678–79.
 Abbott R, ‘The Artificial Inventor Project’ (World Intellectual Property
Organization) <[Link] accessed on 11/09/2021.
 Yu R, ‘Should an Artificial Intelligence be allowed to Get a Patent?’ (Robohub, 9
March 2017) <[Link]
patent/.> accessed on 26/09/2021.

30 | P a g e
 Subramaniam B, ‘Patent Office Reboots CRI Guidelines Yet Again: Removes ‘novel’
Hardware Requirement’ (Spicy IP, 5 July 2017) <[Link]
[Link].>
accessed on 26/09/2021.
 Nishith Desai Associates, ‘The Future is here: Artificial Intelligence and Robotics’
(May 2018) <
[Link]
Intelligence_and_Robotics.pdf > accessed on 11/09/2021.

BOOKS
 Merges R & Duffy J, ‘Patent Law and Policy: Cases and Materials’ (5th ed, 2011) 7–
8.
STATUTES
 The Patents Act 1970.
 35 USC, 1952.
CASES
 Bilsk v Kappos 561 U.S. 593 (2010).
 Diamond v Chakrabarty 447 US 303, 308 (1980).
 Funk Brothers Seed Co v Kalo Inoculant Co 333 US 127.
 McRO Inc v Bandai Namco Games America Inc (Fed Cir 2016).
 Telefonaktiebolaget LM Ericsson v Intex Technologies (India) Limited IA No
6735/2014 in CS(OS) No1045/ 2014 (2015) (‘Ericsson’).
 HTC v Apple [2013] EWCA Civ 451.
 New Idea Farm Equip Corp v Sperry Corp 916 F2d 1561, 1566 n4 (Fed Cir 1990).

31 | P a g e
32 | P a g e