Page 12 - The Indian EYE 060322

OPINION JUNE 03, 2022 | The Indian Eye 12

stage. used for overseeing the battlefield ence & Technology Lab (Dstl) on AI ethical considerations of employing
AI has been aggressively used and collecting and archiving signals projects which were £7million for the AI in military systems.
to spread disinformation in the Rus- intelligence. It was stated at the De- year 2021/22 and is supposed to in- In future, AI will act as an en-
sia–Ukraine War. Machine learning fense One’s Genius AI Summit in crease to £29 million in the next year. abler to out-adapt competitors and
algorithms have been used to amplify April 2022 that all this information The influence of AI on NATO adversaries. The current AI strategy
misleading and fake content on so- will be fed into systems for train- comes with a set of opportunities, of NATO needs to address the vul-
cial media platforms, like doctored ing of machine learning algorithms challenges and risks. Its adoption nerabilities in AI systems and related
videos of invading forces and fake to support future decision-making process has been incremental and measures for effectively using auton-
live streams. Russia is said to have processes. It is believed that the US prescriptive. The rising geopolitical omous weapon systems and military
used AI-enabled systems not only on and NATO allies have already built conflicts and the use of AI in such governance of AI. The NATO accel-
the battlefield but also in cyberspace, such AI-enabled cyber weapons and conflicts have required the establish- erator has been devised to address,
targeting the critical infrastructures defenses, information about which is ment of a dynamic ecosystem to sup- prioritise, and promote interopera-
of Ukraine. Russian troll farms have said to be highly classified. port interoperability. The military bility in transatlantic cooperation to
been alleged to have used AI-en- The vulnerabilities in the AI sys- adoption of AI requires an innova- drive the strategic innovation process.
abled systems to generate human tems can include data poisoning and tion ecosystem that is self-sufficient, Furthermore, NATO needs to
faces for fake propagandist personas input attacks, attacking the supply supports deterrence and resilience, protect the use of AI from manipula-
on social media platforms like Twit- pipelines by simply crafting data and and encompasses the strategic inno- tion and disruption and align it with
ter, Instagram and Facebook. feeding it to public resources, white- vation process. its stated principle of “Responsible
Similarly, Ukraine has been giv- box and black-box attacks. There NATO’s AI strategy raises many use of AI”. NATO needs to work on
en free access to Clearview AI facial is always a chance of orchestrated concerns related to the AI-driven au- AI adoption challenges centred on in-
recognition software, which has a and conflicting data in the face of AI tonomous weapon systems, as it does novation and arms control. It can look
database of 2 billion photos crawled models to derail them and to exploit not adequately address the develop- towards bringing in guiding principles
from Russian social media platforms. the vulnerabilities in the algorithms, ment of such systems, its deployment on use of AI-driven lethal autono-
This software is being used for the and active manipulation by the ad- and governance. mous weapon systems. It is expected
detection of Russian forces and to versaries can be induced. Another challenge for NATO that in the next 2–3 years, AI’s use will
identify the dead and gauge the Defense Advanced Research is to standardize rules for all mem- be confined to the field of military lo-
spread of disinformation in cyber- Projects Agency (DARPA) has ber states in dealing with AI-en- gistics, reconnaissance, mission plan-
space. AI’s analytical potential has launched a Guaranteeing AI Ro- abled autonomous weapon systems. ning and support, predictive main-
been tapped by companies even be- bustness against Deception (GARD) Countries like Turkey are working tenance of a military facility, data
fore the Russia–Ukraine war started. programme. Under this programme, on autonomous weapons and have fusion and analysis, cyber defense
The Russia–Ukraine conflict has development efforts are being made developed AI-enabled loitering mu- and optimization of processes. In the
become a test case for AI adoption to establish a theoretical foundation nitions. Turkey has requested the US long run, NATO could employ AI for
in modern warfare. The US is using for defensible ML and the creation for upgraded F-16 fighter jets that more complex military applications
the conflict as a test-bed for many of and testing of such systems. are said to be AI-enabled. as it generates greater political sup-
its AI projects with the Pentagon’s AI technology in modern war- NATO will also have to focus port for offensive AI military projects.
‘Maven’ project having contributed fare will be an intractable weapon on the vulnerabilities and intrusion
to the detection and classification in future conflicts beyond Ukraine. issues with the AI-enabled systems Dr Sanur Sharma is Associate Fellow
of objects of interest from various Countries trying to achieve a techno- and will need to set up dedicated at Manohar Parrikar Institute for
drone footage through AI and Ma- logical edge over others have started centers for AI development and test- Defense Studies and Analyses
chine Learning (ML) algorithms. It considerable investments in AI tech- ing in order to maintain a test-safe- Views expressed are of the author and
has been reported that the Pentagon nology to strengthen their militaries. ty regime for systems-of-systems do not necessarily reflect the views of
has been using AI and ML tools to NATO has invested US$ 1 billion to employed using AI. The challenges the Manohar Parrikar IDSA or of the
collect a vast amount of data on the develop new AI defense technolo- related to AI use in wars and geopo- Government of India.
Russia–Ukraine war and analyses gies. The US DoD has also planned litical conflicts need to be addressed
it to learn and generate battlefield to invest US$ 874 million in AI-re- to generate confidence in the use of This is the abridged version of the
intelligence about the Russian com- lated technologies as a part of their such systems. Additionally, testing article which appeared first in the
mand and control strategies. army research and development mechanisms and accuracy standards Comment section of the website (www.
The advanced AI-enabled sys- budget (federal fiscal year 2022 DoD need to be implemented for system idsa.in) of Manohar Parrikar Institute
tems with the US Department of De- budget). The UK DoD is funding components. Policymakers need to for Defense Studies and Analyses, New
fense (DoD) are said to have been suppliers to work with Defense Sci- address the operational risks and Delhi on May 24, 2022

www.TheIndianEYE.com