Yojana Magazine November 2023 - 3 | Current Affairs & Hindu Analysis: Daily, Weekly & Monthly - UPSC PDF Download

 Page 1


47 November 2023
Re SPONSibLe USe OF Ai 
bRidGiNG iNNOv AtiON ANd e thiCS
rtificial i ntelligence (Ai) is transforming 
the way humans interact, industries 
function, and societies are structured. 
t he seemingly limitless potential of Ai 
across multiple domains, countries, and human 
imaginations has spawned numerous applications. 
current applications include image and text 
analysis for data analysis purposes, logistics, 
assistance in decision-making, autonomous 
vehicles, and aerial systems, cybersecurity, etc. 
The New Delhi Leaders’ Declaration highlights the significance of 
harnessing ‘ AI responsibly for good and for all’ . It states that the G20 
leaders are committed to leveraging AI for the public good by solving 
challenges in a responsible, inclusive, and human-centric manner while 
protecting people’s rights and safety. Groupings like these are in an 
opportune position to take the lead in this regard, thereby bridging the 
gap between innovation and the ethics of the use of AI.
Additionally, it is being used for security, 
surveillance, and inventory management. it is also 
being applied extensively to areas like agriculture, 
fintech, healthcare, manufacturing, and climate 
change, yielding sizeable dividends in all of them. 
it has become abundantly clear in the recent 
past that Ai can augment human capabilities 
and aid us in tackling some of the most pressing 
challenges of our time. Ai is a force that has the 
ShImona mohan
dR SameeR paTIL 
 The co-author is a Junior Fellow at the Centre for Security, Strategy and Technology, ORF. She works at the intersection of security, technology 
(especially AI and cybersecurity), gender and disarmament. Email: shimona.mohan@orfonline.org.
The author is a Senior Fellow and Deputy Director at the Observer Research Foundation (ORF). He works at the intersection of technology and 
national security. Email: sameer.patil@orfonline.org. 
A
Page 2


47 November 2023
Re SPONSibLe USe OF Ai 
bRidGiNG iNNOv AtiON ANd e thiCS
rtificial i ntelligence (Ai) is transforming 
the way humans interact, industries 
function, and societies are structured. 
t he seemingly limitless potential of Ai 
across multiple domains, countries, and human 
imaginations has spawned numerous applications. 
current applications include image and text 
analysis for data analysis purposes, logistics, 
assistance in decision-making, autonomous 
vehicles, and aerial systems, cybersecurity, etc. 
The New Delhi Leaders’ Declaration highlights the significance of 
harnessing ‘ AI responsibly for good and for all’ . It states that the G20 
leaders are committed to leveraging AI for the public good by solving 
challenges in a responsible, inclusive, and human-centric manner while 
protecting people’s rights and safety. Groupings like these are in an 
opportune position to take the lead in this regard, thereby bridging the 
gap between innovation and the ethics of the use of AI.
Additionally, it is being used for security, 
surveillance, and inventory management. it is also 
being applied extensively to areas like agriculture, 
fintech, healthcare, manufacturing, and climate 
change, yielding sizeable dividends in all of them. 
it has become abundantly clear in the recent 
past that Ai can augment human capabilities 
and aid us in tackling some of the most pressing 
challenges of our time. Ai is a force that has the 
ShImona mohan
dR SameeR paTIL 
 The co-author is a Junior Fellow at the Centre for Security, Strategy and Technology, ORF. She works at the intersection of security, technology 
(especially AI and cybersecurity), gender and disarmament. Email: shimona.mohan@orfonline.org.
The author is a Senior Fellow and Deputy Director at the Observer Research Foundation (ORF). He works at the intersection of technology and 
national security. Email: sameer.patil@orfonline.org. 
A
48 November 2023
capacity to create a more sustainable, equitable, 
and interconnected world. However, it also raises 
critical ethical and societal concerns, which require 
adequate policy consideration and responses. 
this highlights the need for the responsible 
development and deployment of Ai to ensure that 
its transformative power benefits everyone and 
leaves no one behind.
G20 new delhi leaders’ declaration and 
responsible ai
states are increasingly being compelled 
to practise responsible behaviour in their 
engagements with Ai for civilian, security and 
defence purposes. in this context, the recently 
concluded g20 summit in new Delhi (9-10 
september 2023) has tackled multiple aspects 
related to r esponsible Ai (r Ai). Most of the g20 
members have been working towards establishing 
regulations for the responsible use of Ai, especially 
since the advent of genAi applications. the 
european union’s proposed Ai Act is the most 
comprehensive attempt to establish a regulatory 
framework for the responsible development 
of Ai that focuses primarily on strengthening 
rules around data quality, transparency, human 
oversight, and accountability.
1
t he new Delhi Leaders’ Declaration highlights 
the significance of harnessing ‘Ai responsibly for 
good and for all’ .
2
 it states that the g20 leaders are 
committed to leveraging Ai for the public good 
by solving challenges in a responsible, inclusive, 
and human-centric manner, while protecting 
people’s rights and safety. it adds that to ensure 
responsible Ai development, deployment and use, 
the protection of human rights, transparency and 
explainability, fairness, accountability, regulation, 
safety, appropriate human oversight, ethics, biases, 
privacy, and data protection must be addressed. 
in addition, the declaration mentions that the g20 
members will pursue a pro-innovation regulatory/
governance approach that maximises the benefits 
and takes into account the risks associated with 
the use of Ai. 
the declaration also reaffirms the leaders’ 
commitment to g20 Ai Principles of 2019. 
these principles had been adopted at the 
2019 osaka summit and underline the human-
centred approach of Ai.
3
 they take a cue from 
the organisation for economic cooperation and 
Development principles on Ai, also adopted in 
2019, that support the technology to become 
innovative and trustworthy, and respect human 
rights and democratic values.
4
 Besides this, the 
declaration also underlines the importance 
of investment in supporting human capital 
development. towards this, g20 leaders agreed 
to extend support to educational institutions 
and teachers to enable them to keep pace with 
emerging trends and technological advances 
including Ai. this will play an important role in 
imparting skills for the youth entering the job 
market and will offset the concerns around the 
adverse economic impacts of Ai.
h ow does ai pose e thical r isks?
According to the AiAAic (Ai, Algorithmic, 
and Automation incidents and controversies) 
database, which tracks incidents related to the 
ethical misuse of Ai, the number of Ai incidents 
and controversies has increased 26 times since 
2012.
5
several critics of Ai have also raised concerns 
about gender and racial bias when it comes to the 
application of Ai to services like healthcare and 
finance. Although it may appear to be so, Ai is not 
neutral; it can internalise and then catastrophically 
enhance biases that societies possess, programme 
them into the code, and/or ignore them in outputs 
Page 3


47 November 2023
Re SPONSibLe USe OF Ai 
bRidGiNG iNNOv AtiON ANd e thiCS
rtificial i ntelligence (Ai) is transforming 
the way humans interact, industries 
function, and societies are structured. 
t he seemingly limitless potential of Ai 
across multiple domains, countries, and human 
imaginations has spawned numerous applications. 
current applications include image and text 
analysis for data analysis purposes, logistics, 
assistance in decision-making, autonomous 
vehicles, and aerial systems, cybersecurity, etc. 
The New Delhi Leaders’ Declaration highlights the significance of 
harnessing ‘ AI responsibly for good and for all’ . It states that the G20 
leaders are committed to leveraging AI for the public good by solving 
challenges in a responsible, inclusive, and human-centric manner while 
protecting people’s rights and safety. Groupings like these are in an 
opportune position to take the lead in this regard, thereby bridging the 
gap between innovation and the ethics of the use of AI.
Additionally, it is being used for security, 
surveillance, and inventory management. it is also 
being applied extensively to areas like agriculture, 
fintech, healthcare, manufacturing, and climate 
change, yielding sizeable dividends in all of them. 
it has become abundantly clear in the recent 
past that Ai can augment human capabilities 
and aid us in tackling some of the most pressing 
challenges of our time. Ai is a force that has the 
ShImona mohan
dR SameeR paTIL 
 The co-author is a Junior Fellow at the Centre for Security, Strategy and Technology, ORF. She works at the intersection of security, technology 
(especially AI and cybersecurity), gender and disarmament. Email: shimona.mohan@orfonline.org.
The author is a Senior Fellow and Deputy Director at the Observer Research Foundation (ORF). He works at the intersection of technology and 
national security. Email: sameer.patil@orfonline.org. 
A
48 November 2023
capacity to create a more sustainable, equitable, 
and interconnected world. However, it also raises 
critical ethical and societal concerns, which require 
adequate policy consideration and responses. 
this highlights the need for the responsible 
development and deployment of Ai to ensure that 
its transformative power benefits everyone and 
leaves no one behind.
G20 new delhi leaders’ declaration and 
responsible ai
states are increasingly being compelled 
to practise responsible behaviour in their 
engagements with Ai for civilian, security and 
defence purposes. in this context, the recently 
concluded g20 summit in new Delhi (9-10 
september 2023) has tackled multiple aspects 
related to r esponsible Ai (r Ai). Most of the g20 
members have been working towards establishing 
regulations for the responsible use of Ai, especially 
since the advent of genAi applications. the 
european union’s proposed Ai Act is the most 
comprehensive attempt to establish a regulatory 
framework for the responsible development 
of Ai that focuses primarily on strengthening 
rules around data quality, transparency, human 
oversight, and accountability.
1
t he new Delhi Leaders’ Declaration highlights 
the significance of harnessing ‘Ai responsibly for 
good and for all’ .
2
 it states that the g20 leaders are 
committed to leveraging Ai for the public good 
by solving challenges in a responsible, inclusive, 
and human-centric manner, while protecting 
people’s rights and safety. it adds that to ensure 
responsible Ai development, deployment and use, 
the protection of human rights, transparency and 
explainability, fairness, accountability, regulation, 
safety, appropriate human oversight, ethics, biases, 
privacy, and data protection must be addressed. 
in addition, the declaration mentions that the g20 
members will pursue a pro-innovation regulatory/
governance approach that maximises the benefits 
and takes into account the risks associated with 
the use of Ai. 
the declaration also reaffirms the leaders’ 
commitment to g20 Ai Principles of 2019. 
these principles had been adopted at the 
2019 osaka summit and underline the human-
centred approach of Ai.
3
 they take a cue from 
the organisation for economic cooperation and 
Development principles on Ai, also adopted in 
2019, that support the technology to become 
innovative and trustworthy, and respect human 
rights and democratic values.
4
 Besides this, the 
declaration also underlines the importance 
of investment in supporting human capital 
development. towards this, g20 leaders agreed 
to extend support to educational institutions 
and teachers to enable them to keep pace with 
emerging trends and technological advances 
including Ai. this will play an important role in 
imparting skills for the youth entering the job 
market and will offset the concerns around the 
adverse economic impacts of Ai.
h ow does ai pose e thical r isks?
According to the AiAAic (Ai, Algorithmic, 
and Automation incidents and controversies) 
database, which tracks incidents related to the 
ethical misuse of Ai, the number of Ai incidents 
and controversies has increased 26 times since 
2012.
5
several critics of Ai have also raised concerns 
about gender and racial bias when it comes to the 
application of Ai to services like healthcare and 
finance. Although it may appear to be so, Ai is not 
neutral; it can internalise and then catastrophically 
enhance biases that societies possess, programme 
them into the code, and/or ignore them in outputs 
49 November 2023
in the absence of sensitivities to those biases, to 
begin with.
6
 if the datasets used in developing 
any Ai system are incomplete or skewed towards 
or against a sub-group, they will produce results 
that marginalise those sub-groups or make them 
invisible in some way. Y et, even if a dataset is precise 
and representative of the intended population, 
biased Machine Learning (ML) algorithms applied 
to the data may still result in biased outputs. 
in most supervised ML models, training 
datasets are given labels by a human developer 
or coder to enable the ML model to classify 
the information it already has. the model then 
characterises new information given to it based on 
this classification syntax, after which it generates 
an output. t here are two possible modes of bias 
introduction in this process: first, if the human 
developers have their own biases, which they either 
introduce into the system or retain due to ignorant 
oversight; and second, if biases are incorporated 
in the processing of the data within the ‘black box’ 
of the Ai/ML system, that is not explainable to or 
understandable by human operators.
7
 the black 
box, as the name suggests, makes the learning 
process of the system opaque, and its algorithms 
can thus only be fixed once an output is generated 
and the human developer affirms that there was a 
problem with processing the input data.
Besides this, there are also ethical concerns that 
have arisen over issues like copyright infringement 
and privacy violations due to apps that create 
realistic images and art from a description in 
natural language.
8,9
 several artists have accused 
apps of training their algorithms based on images 
and illustrations scraped from the web without 
the original artists’ consent.
10
then there are concerns regarding the 
misuse of Ai in the defence domain to enhance 
targeting and surveillance capabilities of drones 
on the battlefield. t his is a use-case of Ai in drone 
warfare with the potential of ensuing violence.  
in other cases, critics have also noted the misuse 
of Ai for illegal surveillance. in the cybersecurity 
sphere, generative Ai applications are increasingly 
posing legitimate security threats as they are 
being used to conduct malware attacks. For 
instance, cybercriminals, with the help of Ai, mass 
generating phishing emails to spread malware 
and collect valuable information. these phishing 
emails have higher rates of success, than manually 
crafted phishing emails. However, an even more 
insidious threat has emerged through ‘deepfakes,’ 
which generate synthetic or artificial media using 
ML. such realistic-looking content is difficult 
to verify and have become a powerful tool for 
disinformation, with grave national security 
implications. For instance, in March 2022, a deep 
fake video of ukrainian President volodymyr 
Zelenskyy asking his troops to surrender went 
viral among ukrainian citizens, causing significant 
confusion, even as their military was fighting 
against the russian forces.
11
Beyond defence and security, Ai has also 
evoked fears of adverse economic impact. An 
emerging apprehension is that Ai automation 
could potentially alter the labour market in a 
fundamental manner, with grave implications for 
economies in the global s outh that rely on their 
labour and human resources.
12,13
What is responsible ai?
these dynamics have created the necessity 
for the ‘responsible Ai’ (rAi) and the need to 
regulate it. t here has been a gradual momentum 
around rallying for responsible innovation 
ecosystems. this is especially valid in the 
development and deployment of Ai, where 
there is a chance for responsible innovation and 
use to be institutionalised right from the get-
go and not as an afterthought or a checkbox to 
performatively satisfy policy and/or compliance-
Page 4


47 November 2023
Re SPONSibLe USe OF Ai 
bRidGiNG iNNOv AtiON ANd e thiCS
rtificial i ntelligence (Ai) is transforming 
the way humans interact, industries 
function, and societies are structured. 
t he seemingly limitless potential of Ai 
across multiple domains, countries, and human 
imaginations has spawned numerous applications. 
current applications include image and text 
analysis for data analysis purposes, logistics, 
assistance in decision-making, autonomous 
vehicles, and aerial systems, cybersecurity, etc. 
The New Delhi Leaders’ Declaration highlights the significance of 
harnessing ‘ AI responsibly for good and for all’ . It states that the G20 
leaders are committed to leveraging AI for the public good by solving 
challenges in a responsible, inclusive, and human-centric manner while 
protecting people’s rights and safety. Groupings like these are in an 
opportune position to take the lead in this regard, thereby bridging the 
gap between innovation and the ethics of the use of AI.
Additionally, it is being used for security, 
surveillance, and inventory management. it is also 
being applied extensively to areas like agriculture, 
fintech, healthcare, manufacturing, and climate 
change, yielding sizeable dividends in all of them. 
it has become abundantly clear in the recent 
past that Ai can augment human capabilities 
and aid us in tackling some of the most pressing 
challenges of our time. Ai is a force that has the 
ShImona mohan
dR SameeR paTIL 
 The co-author is a Junior Fellow at the Centre for Security, Strategy and Technology, ORF. She works at the intersection of security, technology 
(especially AI and cybersecurity), gender and disarmament. Email: shimona.mohan@orfonline.org.
The author is a Senior Fellow and Deputy Director at the Observer Research Foundation (ORF). He works at the intersection of technology and 
national security. Email: sameer.patil@orfonline.org. 
A
48 November 2023
capacity to create a more sustainable, equitable, 
and interconnected world. However, it also raises 
critical ethical and societal concerns, which require 
adequate policy consideration and responses. 
this highlights the need for the responsible 
development and deployment of Ai to ensure that 
its transformative power benefits everyone and 
leaves no one behind.
G20 new delhi leaders’ declaration and 
responsible ai
states are increasingly being compelled 
to practise responsible behaviour in their 
engagements with Ai for civilian, security and 
defence purposes. in this context, the recently 
concluded g20 summit in new Delhi (9-10 
september 2023) has tackled multiple aspects 
related to r esponsible Ai (r Ai). Most of the g20 
members have been working towards establishing 
regulations for the responsible use of Ai, especially 
since the advent of genAi applications. the 
european union’s proposed Ai Act is the most 
comprehensive attempt to establish a regulatory 
framework for the responsible development 
of Ai that focuses primarily on strengthening 
rules around data quality, transparency, human 
oversight, and accountability.
1
t he new Delhi Leaders’ Declaration highlights 
the significance of harnessing ‘Ai responsibly for 
good and for all’ .
2
 it states that the g20 leaders are 
committed to leveraging Ai for the public good 
by solving challenges in a responsible, inclusive, 
and human-centric manner, while protecting 
people’s rights and safety. it adds that to ensure 
responsible Ai development, deployment and use, 
the protection of human rights, transparency and 
explainability, fairness, accountability, regulation, 
safety, appropriate human oversight, ethics, biases, 
privacy, and data protection must be addressed. 
in addition, the declaration mentions that the g20 
members will pursue a pro-innovation regulatory/
governance approach that maximises the benefits 
and takes into account the risks associated with 
the use of Ai. 
the declaration also reaffirms the leaders’ 
commitment to g20 Ai Principles of 2019. 
these principles had been adopted at the 
2019 osaka summit and underline the human-
centred approach of Ai.
3
 they take a cue from 
the organisation for economic cooperation and 
Development principles on Ai, also adopted in 
2019, that support the technology to become 
innovative and trustworthy, and respect human 
rights and democratic values.
4
 Besides this, the 
declaration also underlines the importance 
of investment in supporting human capital 
development. towards this, g20 leaders agreed 
to extend support to educational institutions 
and teachers to enable them to keep pace with 
emerging trends and technological advances 
including Ai. this will play an important role in 
imparting skills for the youth entering the job 
market and will offset the concerns around the 
adverse economic impacts of Ai.
h ow does ai pose e thical r isks?
According to the AiAAic (Ai, Algorithmic, 
and Automation incidents and controversies) 
database, which tracks incidents related to the 
ethical misuse of Ai, the number of Ai incidents 
and controversies has increased 26 times since 
2012.
5
several critics of Ai have also raised concerns 
about gender and racial bias when it comes to the 
application of Ai to services like healthcare and 
finance. Although it may appear to be so, Ai is not 
neutral; it can internalise and then catastrophically 
enhance biases that societies possess, programme 
them into the code, and/or ignore them in outputs 
49 November 2023
in the absence of sensitivities to those biases, to 
begin with.
6
 if the datasets used in developing 
any Ai system are incomplete or skewed towards 
or against a sub-group, they will produce results 
that marginalise those sub-groups or make them 
invisible in some way. Y et, even if a dataset is precise 
and representative of the intended population, 
biased Machine Learning (ML) algorithms applied 
to the data may still result in biased outputs. 
in most supervised ML models, training 
datasets are given labels by a human developer 
or coder to enable the ML model to classify 
the information it already has. the model then 
characterises new information given to it based on 
this classification syntax, after which it generates 
an output. t here are two possible modes of bias 
introduction in this process: first, if the human 
developers have their own biases, which they either 
introduce into the system or retain due to ignorant 
oversight; and second, if biases are incorporated 
in the processing of the data within the ‘black box’ 
of the Ai/ML system, that is not explainable to or 
understandable by human operators.
7
 the black 
box, as the name suggests, makes the learning 
process of the system opaque, and its algorithms 
can thus only be fixed once an output is generated 
and the human developer affirms that there was a 
problem with processing the input data.
Besides this, there are also ethical concerns that 
have arisen over issues like copyright infringement 
and privacy violations due to apps that create 
realistic images and art from a description in 
natural language.
8,9
 several artists have accused 
apps of training their algorithms based on images 
and illustrations scraped from the web without 
the original artists’ consent.
10
then there are concerns regarding the 
misuse of Ai in the defence domain to enhance 
targeting and surveillance capabilities of drones 
on the battlefield. t his is a use-case of Ai in drone 
warfare with the potential of ensuing violence.  
in other cases, critics have also noted the misuse 
of Ai for illegal surveillance. in the cybersecurity 
sphere, generative Ai applications are increasingly 
posing legitimate security threats as they are 
being used to conduct malware attacks. For 
instance, cybercriminals, with the help of Ai, mass 
generating phishing emails to spread malware 
and collect valuable information. these phishing 
emails have higher rates of success, than manually 
crafted phishing emails. However, an even more 
insidious threat has emerged through ‘deepfakes,’ 
which generate synthetic or artificial media using 
ML. such realistic-looking content is difficult 
to verify and have become a powerful tool for 
disinformation, with grave national security 
implications. For instance, in March 2022, a deep 
fake video of ukrainian President volodymyr 
Zelenskyy asking his troops to surrender went 
viral among ukrainian citizens, causing significant 
confusion, even as their military was fighting 
against the russian forces.
11
Beyond defence and security, Ai has also 
evoked fears of adverse economic impact. An 
emerging apprehension is that Ai automation 
could potentially alter the labour market in a 
fundamental manner, with grave implications for 
economies in the global s outh that rely on their 
labour and human resources.
12,13
What is responsible ai?
these dynamics have created the necessity 
for the ‘responsible Ai’ (rAi) and the need to 
regulate it. t here has been a gradual momentum 
around rallying for responsible innovation 
ecosystems. this is especially valid in the 
development and deployment of Ai, where 
there is a chance for responsible innovation and 
use to be institutionalised right from the get-
go and not as an afterthought or a checkbox to 
performatively satisfy policy and/or compliance-
50 November 2023
related constraints. in this context, rAi is 
broadly understood as the practice of designing, 
developing, and deploying Ai to empower 
employees and businesses and impact society in 
a fair manner. given Ai’s dual-use character, this is 
a loose and flexible understanding, and it posits 
r Ai as an umbrella term that usually encompasses 
considerations around fair, explainable, and 
trustworthy Ai systems.
india has been working on rAi since 
2018, and niti Aayog also released a two-part 
report in 2021 on approaches towards
14
 and 
operationalisation of
15
 rAi principles for the 
deployment and use of civilian Ai architectures. 
t he seven principles that niti Aayog highlights 
are: safety and reliability; equality; inclusivity 
and non-discrimination; privacy and security; 
transparency; accountability; and protection 
and reinforcement of positive human values. it 
also recommends measures for the government, 
industry bodies, and civil society to implement 
these principles in the Ai products they develop 
or work with. indian tech industry body 
nAsscoM embedded the principles of this 
framework into india’s first r Ai Hub and t oolkit
16
 
released in late 2022, which comprises sector-
agnostic tools to enable entities to leverage Ai 
by prioritising user trust and safety.
Pertinently, the focus on r Ai in g20 new Delhi 
Leaders’ Declaration also aligns with india holding 
the chair of the global Partnership on Artificial 
intelligence (gPAi), a multistakeholder initiative 
that brings together experts from science, industry, 
civil society, international organisations, and 
governments.
17
 it contributes to the responsible 
development of Ai via its r esponsible Ai working 
group.
18
india chairing the gPAi is important 
since the global south is underrepresented in 
the forum: out of its 29 members, only four are 
from the global south - Argentina, Brazil, india, 
and senegal. t herefore, india is better positioned 
to play an active role in bridging this divide and 
ensuring that the less developed economies also 
get to reap the benefits of this technological shift 
towards Ai. new Delhi will host the annual gPAi 
summit on 12-14 December 2023. At the last 
year’s summit in tokyo, india urged the members 
to work together on a common framework of rules 
and guidelines on data governance in order to 
prevent user harm and ensure the safety of both 
the internet and Ai.
c onclusion 
though the rise of Ai and its applications in 
the past few years has been meteoric and the 
scope for innovation in the field is endless, nations 
all around the world are waking up to the dangers 
of its potential misuse. While there are several 
initiatives attempting to address the issue, there 
is currently no global consensus or regulatory 
framework on the ethical and responsible use 
of Ai. Hence, groupings like the g20 and gPAi 
are in an opportune position to take the lead in 
this regard, thereby bridging the gap between 
innovation and the ethics of Ai use. the g20 
new Delhi Leaders’ Declaration demonstrates 
that leaders of the world’s largest economies are 
aware of the potential benefits and risks of Ai and 
are committed to working together to ensure 
that the technology is developed and used in 
a responsible and inclusive manner. the g20 
members must follow this declaration by adopting 
the anticipatory regulation approach, doing over-
the-horizon thinking, and building a coalition of 
diverse stakeholders. ?
Page 5


47 November 2023
Re SPONSibLe USe OF Ai 
bRidGiNG iNNOv AtiON ANd e thiCS
rtificial i ntelligence (Ai) is transforming 
the way humans interact, industries 
function, and societies are structured. 
t he seemingly limitless potential of Ai 
across multiple domains, countries, and human 
imaginations has spawned numerous applications. 
current applications include image and text 
analysis for data analysis purposes, logistics, 
assistance in decision-making, autonomous 
vehicles, and aerial systems, cybersecurity, etc. 
The New Delhi Leaders’ Declaration highlights the significance of 
harnessing ‘ AI responsibly for good and for all’ . It states that the G20 
leaders are committed to leveraging AI for the public good by solving 
challenges in a responsible, inclusive, and human-centric manner while 
protecting people’s rights and safety. Groupings like these are in an 
opportune position to take the lead in this regard, thereby bridging the 
gap between innovation and the ethics of the use of AI.
Additionally, it is being used for security, 
surveillance, and inventory management. it is also 
being applied extensively to areas like agriculture, 
fintech, healthcare, manufacturing, and climate 
change, yielding sizeable dividends in all of them. 
it has become abundantly clear in the recent 
past that Ai can augment human capabilities 
and aid us in tackling some of the most pressing 
challenges of our time. Ai is a force that has the 
ShImona mohan
dR SameeR paTIL 
 The co-author is a Junior Fellow at the Centre for Security, Strategy and Technology, ORF. She works at the intersection of security, technology 
(especially AI and cybersecurity), gender and disarmament. Email: shimona.mohan@orfonline.org.
The author is a Senior Fellow and Deputy Director at the Observer Research Foundation (ORF). He works at the intersection of technology and 
national security. Email: sameer.patil@orfonline.org. 
A
48 November 2023
capacity to create a more sustainable, equitable, 
and interconnected world. However, it also raises 
critical ethical and societal concerns, which require 
adequate policy consideration and responses. 
this highlights the need for the responsible 
development and deployment of Ai to ensure that 
its transformative power benefits everyone and 
leaves no one behind.
G20 new delhi leaders’ declaration and 
responsible ai
states are increasingly being compelled 
to practise responsible behaviour in their 
engagements with Ai for civilian, security and 
defence purposes. in this context, the recently 
concluded g20 summit in new Delhi (9-10 
september 2023) has tackled multiple aspects 
related to r esponsible Ai (r Ai). Most of the g20 
members have been working towards establishing 
regulations for the responsible use of Ai, especially 
since the advent of genAi applications. the 
european union’s proposed Ai Act is the most 
comprehensive attempt to establish a regulatory 
framework for the responsible development 
of Ai that focuses primarily on strengthening 
rules around data quality, transparency, human 
oversight, and accountability.
1
t he new Delhi Leaders’ Declaration highlights 
the significance of harnessing ‘Ai responsibly for 
good and for all’ .
2
 it states that the g20 leaders are 
committed to leveraging Ai for the public good 
by solving challenges in a responsible, inclusive, 
and human-centric manner, while protecting 
people’s rights and safety. it adds that to ensure 
responsible Ai development, deployment and use, 
the protection of human rights, transparency and 
explainability, fairness, accountability, regulation, 
safety, appropriate human oversight, ethics, biases, 
privacy, and data protection must be addressed. 
in addition, the declaration mentions that the g20 
members will pursue a pro-innovation regulatory/
governance approach that maximises the benefits 
and takes into account the risks associated with 
the use of Ai. 
the declaration also reaffirms the leaders’ 
commitment to g20 Ai Principles of 2019. 
these principles had been adopted at the 
2019 osaka summit and underline the human-
centred approach of Ai.
3
 they take a cue from 
the organisation for economic cooperation and 
Development principles on Ai, also adopted in 
2019, that support the technology to become 
innovative and trustworthy, and respect human 
rights and democratic values.
4
 Besides this, the 
declaration also underlines the importance 
of investment in supporting human capital 
development. towards this, g20 leaders agreed 
to extend support to educational institutions 
and teachers to enable them to keep pace with 
emerging trends and technological advances 
including Ai. this will play an important role in 
imparting skills for the youth entering the job 
market and will offset the concerns around the 
adverse economic impacts of Ai.
h ow does ai pose e thical r isks?
According to the AiAAic (Ai, Algorithmic, 
and Automation incidents and controversies) 
database, which tracks incidents related to the 
ethical misuse of Ai, the number of Ai incidents 
and controversies has increased 26 times since 
2012.
5
several critics of Ai have also raised concerns 
about gender and racial bias when it comes to the 
application of Ai to services like healthcare and 
finance. Although it may appear to be so, Ai is not 
neutral; it can internalise and then catastrophically 
enhance biases that societies possess, programme 
them into the code, and/or ignore them in outputs 
49 November 2023
in the absence of sensitivities to those biases, to 
begin with.
6
 if the datasets used in developing 
any Ai system are incomplete or skewed towards 
or against a sub-group, they will produce results 
that marginalise those sub-groups or make them 
invisible in some way. Y et, even if a dataset is precise 
and representative of the intended population, 
biased Machine Learning (ML) algorithms applied 
to the data may still result in biased outputs. 
in most supervised ML models, training 
datasets are given labels by a human developer 
or coder to enable the ML model to classify 
the information it already has. the model then 
characterises new information given to it based on 
this classification syntax, after which it generates 
an output. t here are two possible modes of bias 
introduction in this process: first, if the human 
developers have their own biases, which they either 
introduce into the system or retain due to ignorant 
oversight; and second, if biases are incorporated 
in the processing of the data within the ‘black box’ 
of the Ai/ML system, that is not explainable to or 
understandable by human operators.
7
 the black 
box, as the name suggests, makes the learning 
process of the system opaque, and its algorithms 
can thus only be fixed once an output is generated 
and the human developer affirms that there was a 
problem with processing the input data.
Besides this, there are also ethical concerns that 
have arisen over issues like copyright infringement 
and privacy violations due to apps that create 
realistic images and art from a description in 
natural language.
8,9
 several artists have accused 
apps of training their algorithms based on images 
and illustrations scraped from the web without 
the original artists’ consent.
10
then there are concerns regarding the 
misuse of Ai in the defence domain to enhance 
targeting and surveillance capabilities of drones 
on the battlefield. t his is a use-case of Ai in drone 
warfare with the potential of ensuing violence.  
in other cases, critics have also noted the misuse 
of Ai for illegal surveillance. in the cybersecurity 
sphere, generative Ai applications are increasingly 
posing legitimate security threats as they are 
being used to conduct malware attacks. For 
instance, cybercriminals, with the help of Ai, mass 
generating phishing emails to spread malware 
and collect valuable information. these phishing 
emails have higher rates of success, than manually 
crafted phishing emails. However, an even more 
insidious threat has emerged through ‘deepfakes,’ 
which generate synthetic or artificial media using 
ML. such realistic-looking content is difficult 
to verify and have become a powerful tool for 
disinformation, with grave national security 
implications. For instance, in March 2022, a deep 
fake video of ukrainian President volodymyr 
Zelenskyy asking his troops to surrender went 
viral among ukrainian citizens, causing significant 
confusion, even as their military was fighting 
against the russian forces.
11
Beyond defence and security, Ai has also 
evoked fears of adverse economic impact. An 
emerging apprehension is that Ai automation 
could potentially alter the labour market in a 
fundamental manner, with grave implications for 
economies in the global s outh that rely on their 
labour and human resources.
12,13
What is responsible ai?
these dynamics have created the necessity 
for the ‘responsible Ai’ (rAi) and the need to 
regulate it. t here has been a gradual momentum 
around rallying for responsible innovation 
ecosystems. this is especially valid in the 
development and deployment of Ai, where 
there is a chance for responsible innovation and 
use to be institutionalised right from the get-
go and not as an afterthought or a checkbox to 
performatively satisfy policy and/or compliance-
50 November 2023
related constraints. in this context, rAi is 
broadly understood as the practice of designing, 
developing, and deploying Ai to empower 
employees and businesses and impact society in 
a fair manner. given Ai’s dual-use character, this is 
a loose and flexible understanding, and it posits 
r Ai as an umbrella term that usually encompasses 
considerations around fair, explainable, and 
trustworthy Ai systems.
india has been working on rAi since 
2018, and niti Aayog also released a two-part 
report in 2021 on approaches towards
14
 and 
operationalisation of
15
 rAi principles for the 
deployment and use of civilian Ai architectures. 
t he seven principles that niti Aayog highlights 
are: safety and reliability; equality; inclusivity 
and non-discrimination; privacy and security; 
transparency; accountability; and protection 
and reinforcement of positive human values. it 
also recommends measures for the government, 
industry bodies, and civil society to implement 
these principles in the Ai products they develop 
or work with. indian tech industry body 
nAsscoM embedded the principles of this 
framework into india’s first r Ai Hub and t oolkit
16
 
released in late 2022, which comprises sector-
agnostic tools to enable entities to leverage Ai 
by prioritising user trust and safety.
Pertinently, the focus on r Ai in g20 new Delhi 
Leaders’ Declaration also aligns with india holding 
the chair of the global Partnership on Artificial 
intelligence (gPAi), a multistakeholder initiative 
that brings together experts from science, industry, 
civil society, international organisations, and 
governments.
17
 it contributes to the responsible 
development of Ai via its r esponsible Ai working 
group.
18
india chairing the gPAi is important 
since the global south is underrepresented in 
the forum: out of its 29 members, only four are 
from the global south - Argentina, Brazil, india, 
and senegal. t herefore, india is better positioned 
to play an active role in bridging this divide and 
ensuring that the less developed economies also 
get to reap the benefits of this technological shift 
towards Ai. new Delhi will host the annual gPAi 
summit on 12-14 December 2023. At the last 
year’s summit in tokyo, india urged the members 
to work together on a common framework of rules 
and guidelines on data governance in order to 
prevent user harm and ensure the safety of both 
the internet and Ai.
c onclusion 
though the rise of Ai and its applications in 
the past few years has been meteoric and the 
scope for innovation in the field is endless, nations 
all around the world are waking up to the dangers 
of its potential misuse. While there are several 
initiatives attempting to address the issue, there 
is currently no global consensus or regulatory 
framework on the ethical and responsible use 
of Ai. Hence, groupings like the g20 and gPAi 
are in an opportune position to take the lead in 
this regard, thereby bridging the gap between 
innovation and the ethics of Ai use. the g20 
new Delhi Leaders’ Declaration demonstrates 
that leaders of the world’s largest economies are 
aware of the potential benefits and risks of Ai and 
are committed to working together to ensure 
that the technology is developed and used in 
a responsible and inclusive manner. the g20 
members must follow this declaration by adopting 
the anticipatory regulation approach, doing over-
the-horizon thinking, and building a coalition of 
diverse stakeholders. ?
51 November 2023
references 
1. “eu Ai Act: first regulation on artificial intelligence,” June 
14, 2023, https://www.europarl.europa.eu/news/en/
headlines/society/20230601sto93804/eu-ai-act-first-
regulation-on-artificial-intelligence .
2. g20 new Delhi Leaders’ Declaration, s eptember 9-10, 
2023, https://www.g20.org/content/dam/gtwenty/
gtwenty_new/document/g20-new-Delhi-Leaders-
Declaration.pdf.
3. g20 Ai Principles,  https://www.mofa.go.jp/policy/
economy/g20_summit/osaka19/pdf/documents/en/
annex_08.pdf.
4. “oecD Ai Principles overview,” https://oecd.ai/en/ai-
principles.
5. Artificial i ntelligence index report 2023, stanford 
university Human- c entred Artificial i ntelligence, https://
aiindex.stanford.edu/wp-content/uploads/2023/04/
HAi_Ai-index-r eport-2023_cHAPter_3.pdf .
6. shimona Mohan, “Filling the Blanks: Putting gender 
into Military A.i.,” orF issue Brief no. 655, August 2023, 
observer r esearch Foundation, https://www.orfonline.
org/research/filling-the-blanks-putting-gender-into-
military-ai/. 
7. shimona Mohan, “gender-ative Ai: An enduring gender 
bias in generative Ai systems,”observer research 
Foundation, April 27, 2023, https://www.orfonline.org/
expert-speak/gender-ative-ai/. 
8. DALL.e2, https://openai.com/dall-e-2.
9. Midjourney, https://www.midjourney.com/home/.
10. James vincent, “Ai art tools stable Diffusion and 
Midjourney targeted with copyright lawsuit,” The 
Verge, January 16, 2023, https://www.theverge.
com/2023/1/16/23557098/generative-ai-art-copyright-
legal-lawsuit-stable-diffusion-midjourney-deviantart .
11. t he telegraph, “Deepfake video of volodymyr Zelensky 
surrendering surfaces on social media,” March 17, 2022, 
https://www.youtube.com/watch?v=X17yrev5sl4.
12. ian shine and Kate Whiting, “these are the jobs most 
likely to be lost – and created – because of Ai,” World 
Economic Forum, May 4, 2023, https://www.weforum.
org/agenda/2023/05/jobs-lost-created-ai-gpt/.
13. Accenture, “A new era of generative Ai for everyone,” 
https://www.accenture.com/content/dam/accenture/
final/accenture-com/document/Accenture-A-new-era-
of-generative-Ai-for-everyone.pdf.
14. niti Aayog, “resPonsiBLe Ai #AiFor ALL: Approach 
Document for india Part 1 – Principles for r esponsible 
Ai,” February 2021, https://www.niti.gov.in/sites/default/
files/2021-02/r esponsible-Ai-22022021.pdf.
15. niti Aayog, “resPonsiBLe Ai #AiFor ALL: Approach 
Document for india: Part 2 - operationalizing Principles 
for responsible Ai,” August 2021, https://www.niti.
gov.in/sites/default/files/2021-08/Part2-r esponsible-
Ai-12082021.pdf.
16. inDiAai, “nAsscoM launched the r esponsible Ai hub 
and resource kit,” october 11, 2022, https://indiaai.gov.
in/news/nasscom-launched-the-responsible-ai-hub-
and-resource-kit.
17. Prateek tripathi, “india’s chairmanship of the global 
Partnership on Ai,” observer r esearch Foundation, August 
8, 2023,  https://www.orfonline.org/expert-speak/indias-
chairmanship-of-the-global-partnership-on-ai/.
18. the global Partnership on Artificial i ntelligence, 
“Working group on responsible Ai,” https://gpai.ai/
projects/responsible-ai/.
Sales Outlets of Publications Division
New Delhi Soochna Bhawan, CGO Complex, Lodhi Road 110003  011-24365609 
   011-24365610
Navi Mumbai  701, B Wing, 7th Floor, Kendriya Sadan, Belapur 400614  022-27570686
Kolkata 08, Esplanade East 700069  033-22486696 
Chennai 'A' Wing, Rajaji Bhawan, Basant Nagar 600090  044-24917673 
Thiruvananthapuram Press Road, Near Government Press 695001  0471-2330650
Hyderabad 204, II Floor CGO Towers, Kavadiguda, Secunderabad 500080  040-27535383
Bengaluru I Floor, 'F' Wing, Kendriya Sadan, Koramangala 560034  080-25537244
Patna Bihar State Co-operative Building, Ashoka Rajpath 800004  0612-2675823 
Lucknow Hall No 1, II Floor, Kendriya Bhawan, Sector-H, Aliganj 226024 0522-2325455
Ahmedabad 4-C, Neptune T ower , 4th Floor , Nehru Bridge Corner , Ashram Road  380009 079-26588669