1 of 30

�� Maj Gen P K Mallick, VSM (Retd)

USI

New Delhi

08 Apr 2024

Final Presentation on Employment of AI at Tactical Level – Guide’s Remarks

2 of 30

3 of 30

4 of 30

5 of 30

6 of 30

7 of 30

8 of 30

9 of 30

10 of 30

Elephant in the Room – China

11 of 30

12 of 30

13 of 30

14 of 30

15 of 30

16 of 30

AI-enabled Capabilities can Deliver Game-changing Advantages for Defence

  • Accelerated and better decision-making.
  • Enhanced human cognitive and physical performance.
  • Heightened military readiness and operational competence.
  • New systems of design, manufacture and sustainment of military systems.
  • Capability to produce and detect strategic cyberattacks and information operations.
  • Innovative capabilities that can upset delicate military balances.

17 of 30

�AI and Autonomous Systems are Attractive Technologies to Militaries

  • AI can outperform humans in automated military and civilian tasks and in more demanding tasks. AI has the potential to improve the precision of weapons, increase the accuracy of data analysis and provide earlier alerts of attacks.
  • AI can process information much faster than humans. This speed is crucial as AI-enabled technologies become more integrated into military decision-making processes.
  • AI algorithms may be embedded in and enhance automated systems, which carry out tasks where soldiers risk their lives neutralising mines or conducting intelligence, surveillance and reconnaissance in dangerous areas. Autonomous vehicles can perform these tasks without losing human lives.
  • Autonomous systems can act as a force multiplier. It can perform tasks such as patrolling areas for extended periods without needing to rest or consume resources. They can process data much faster than human analysts. This allows military personnel to be freed up to perform other duties.

18 of 30

Non-Technical and Technical Challenges

  • The leading AI companies may be reluctant to partner with the military due to the complexity of the defence acquisition process.
  • Commercial technology companies are also often reluctant to partner with Military due to intellectual property and data rights concerns. Intellectual property is the “lifeblood” of commercial technology companies
  • Non-technical challenges include investment, innovation and workforce challenges. Militaries will need to invest enough capital in research and development. The armed forces need to improve their ability to adapt and integrate technologies from the non-defence commercial sector and to find ways to attract the top AI experts, many of whom are offered higher pay in the private sector.

19 of 30

Opportunities

  • AI technology is good at solving well-defined, narrow problems where the necessary data and feedback are fully available to the system.
  • Can facilitate autonomous operations, lead to more informed military decision-making and will likely increase the speed and scale of military action.
  • Can be used in military scenario planning at all levels of command to determine the requirements for a mission. AI ensures the most advantageous use of resources during a mission.
  • Although humans will be present, their role will be less significant and the technology will make combat “less uncertain and more controllable,” as machines are not subject to the frailties that cloud human judgment, like being “tired, frightened, bored, or angry.”
  • Far superior at solving multiple control problems very quickly due to their ability to process massive amounts of information to detect patterns without suffering fatigue, recognition error, bias or emotional interference.
  • Non-lethal activities such as logistics, maintenance, base operations, veterans’ healthcare, lifesaving battlefield medical assistance and casualty evacuation, personnel management, navigation, communication, cyber-defence, and intelligence analysis can benefit from AI.
  • AI may play an important role in new systems for protecting people and high-value fixed assets and deterring attacks through non-lethal means.
  • Robots are better suited than soldiers for dull, dirty or dangerous missions.
  • Lethal autonomous robots have the exceptional potential to operate at a tempo faster than humans can achieve and to lethally strike even when communications links have been severed.
  • AI-powered intelligence systems provide the ability to integrate and sort through large troves of data from different sources and geographic locations to identify patterns and highlight useful information, significantly improving intelligence analysis.
  • The judgments of autonomous weapons systems will not be affected by emotions like fear or hysteria. The systems will be capable to process more incoming sensory information than soldiers without discarding or distorting it to fit preconceived notions.
  • Between human and robot soldiers, the robots could be more trusted to report ethical infractions they observed than would a team of soldiers who might close ranks. There is an ethical advantages in removing humans from high-stress combat zones in favour of robots.

20 of 30

Challenges

  • Bureaucratic nature of the armed forces, outdated acquisition and contracting processes and a culture that is reluctant to take risks.
  • Hardware-oriented toward ships, planes and tanks. Spending remains concentrated on legacy systems designed for the industrial age and Cold War. It is now trying to leap into a software-intensive enterprise.
  • Managing the large amounts of data required, ensuring data quality and addressing data storage, access, classification and integration issues. technical limitations in existing systems and technologies impede compatibility with how data is captured and processed, making it difficult to implement AI effectively.
  • Effectiveness of AI algorithms is limited by the quality and nature of the data they are trained on. Obtaining large, high-quality datasets containing military information can be challenging due to its sensitive and classified nature. many military ML algorithms may rely on simulated data that may not accurately reflect real-world conditions. training process of ML algorithms can be vulnerable to manipulation through the introduction of false data, a tactic known as "data poisoning.“

21 of 30

Disadvantages of AI systems

  • Enduring necessity for human presence on the battlefield alongside AI systems in some capacity as the principle restraining factor that will keep the technology from upending warfare.
  • Lethal autonomous weapons will inevitably produce errors. These errors will be difficult to correct or prevent from reoccurring. It is difficult to take corrective action without understanding how the weapon system behaves and why. Automation bias could create denial that an error has even occurred.
  • Lethal autonomous weapons may accrue accidental and collateral damage risk causing mass fratricide, with large numbers of weapons turning on friendly forces and civilian casualties. This could be because of hacking, enemy behavioural manipulation, simple malfunctions, unexpected interactions with the environment or software errors.
  • Lethal autonomous weapon’s error would likely repeat with a consistent level of force until some external agent intervene. Human error tends to be idiosyncratic and one-off, given a human operator’s common sense, moral agency and capacity for near real time consequence management.
  • Waging war via robots is unethical. Because it spurs so much moral outrage among the population from whom the country needs support, robot warfare has serious strategic disadvantages.
  • Service members at every level lack the technical education and experience to employ AI.
  • Current capabilities in the military sphere are far from perfect and mistakes may have serious consequences.
  • Inability to multitask. A fairly simple set of tasks is currently impossible for an AI system to accomplish.

22 of 30

Disadvantages of AI systems

  • Armed forces lag far behind the commercial sector in integrating new and disruptive technologies such as AI into its operations
  • Data required for machine learning is currently isolated, disorganised or frequently discarded and the platforms are not connected. The methods for acquiring, developing, and implementing AI are rigid, sequential processes that hinder early and ongoing experimentation and testing,
  • AI systems can struggle to function reliably or correctly if the data inputs change or encounter uncertain or novel situations.
  • Adversarial attacks on ML systems can target the algorithms or data that the system uses. Even minor modifications can lead to malfunction, incorrect conclusions or other unforeseen issues.
  • AI/ML systems cannot typically explain the reasoning behind their decisions, recommendations and actions. The lack of explainability significantly hinders building trust between human and AI teams.
  • AI technology is unpredictable, vulnerable to unique forms of manipulation and challenges human-machine interaction.

23 of 30

24 of 30

Fake News

25 of 30

26 of 30

27 of 30

LAWS

28 of 30

29 of 30

MISC POINTS

  • Fake News, Deep Fake, AI in IW/PSY OPS
  • Use of AI in Gaza, Lavender
  • Cognitive Warfare
  • AI in Cyber Ops
  • LLM/Chat GPT in Warfare
  • Battle Swarm
  • HMI/HMT (Human Machine Interaction/Teaming)
  • Use of AI by bad actors
  • Use of Robots
  • LAWS, Ethical Issues
  • USA looking for bright STEM graduates from India. They have the labs, we have the people.
  • AI Laws in line with Genève Conventions, IHL
  • Regulation of AI
  • Tech companies looking for collaborative engagement rather than a contract to build a platform. Example of PALANTIR

30 of 30

Thank You