158. Cybersecurity Risks of AI Systems – Detailed Explanation With Case Laws

1. Meaning

Cybersecurity risks of AI systems means the dangers created when Artificial Intelligence (AI) systems are attacked, manipulated, misused or used to attack other systems.

AI is increasingly used in:

Smart grids

Electricity demand forecasting

Energy trading

Smart meters

Nuclear facilities

Renewable-energy management

Banking and financial systems

Transport and smart cities

Because AI depends on data, software, networks and algorithms, a cyberattack can create serious legal and economic consequences.

2. Major Cybersecurity Risks

1. Data Poisoning

Attackers may put false or harmful information into AI training data.

Example: If false electricity-demand data is entered, an AI system may incorrectly predict electricity requirements.

2. Hacking of AI Systems

Hackers may obtain unauthorised access and change AI outputs.

Example: An attacker could manipulate an AI-controlled electricity system and cause incorrect power distribution.

3. Adversarial Attacks

An attacker gives specially designed input to confuse an AI system.

Example: A small change in sensor information may cause an AI system to wrongly identify a fault.

4. Privacy Risk

AI systems can process large amounts of personal information.

Smart electricity systems may reveal:

When a person is at home

Electricity-use habits

Appliance usage

Behavioural patterns

Therefore, privacy can be affected.

5. Automated Decisions

AI may automatically make decisions such as detecting electricity theft or identifying suspicious transactions.

The problem arises when the AI makes a wrong decision and there is no human review.

3. Legal Framework in India

Information Technology Act, 2000

The IT Act provides the basic legal framework for various forms of cyber offences and unauthorised access.

Digital Personal Data Protection Act, 2023

Where AI systems process digital personal data, data-protection obligations become important.

Article 21 of Constitution

Article 21 protects life and personal liberty and has been interpreted to include the right to privacy.

Article 14

AI-based government decisions should not be arbitrary or discriminatory.

4. Important Case Laws

1. K.S. Puttaswamy v. Union of India (2017)

The Supreme Court recognised privacy as a fundamental right under the Constitution.

Relevance to AI

AI systems collect and analyse huge quantities of personal data. Therefore:

AI data collection → privacy risk → constitutional protection

AI systems must have a legitimate purpose and appropriate safeguards.

2. K.S. Puttaswamy v. Union of India (2018)

The Court considered privacy, data protection and proportionality in the Aadhaar context.

Relevance

Government use of AI involving personal data should satisfy principles such as:

Legality

Legitimate purpose

Necessity

Proportionality

Safeguards

3. Shreya Singhal v. Union of India (2015)

The Supreme Court examined provisions of the Information Technology Act and constitutional protection of online speech.

Relevance

Digital technologies must operate within constitutional limits. AI systems used for online monitoring or content decisions should therefore be subject to lawful authority and constitutional safeguards.

4. Anuradha Bhasin v. Union of India (2020)

The Supreme Court examined restrictions affecting internet access and emphasised legality and proportionality.

Relevance

Modern AI systems often depend on digital networks. Government restrictions or cybersecurity measures affecting digital systems should have a proper legal basis and must be proportionate.

5. AI in Energy Systems

Cybersecurity becomes especially important when AI controls critical energy infrastructure.

For example:

AI system → Smart grid → Electricity distribution

If hackers manipulate the AI system, consequences may include:

Wrong electricity forecasting

Grid instability

Incorrect billing

Electricity shortages

Damage to infrastructure

Consumer financial loss

Therefore, AI cybersecurity is also an energy-security issue.

6. Responsibility for AI Cyberattacks

A difficult question is:

Who is legally responsible if an AI system is hacked?

Possible parties include:

AI developer

Energy company

Software provider

Cloud-service provider

Cybersecurity contractor

Government regulator

System operator

Responsibility depends on the contract, applicable legislation, negligence, security obligations and actual cause of the damage.

7. Important Safeguards

AI systems should use:

Strong authentication

Encryption

Regular security testing

Continuous monitoring

Secure software development

Data-quality controls

Human supervision

Incident-response plans

Audit trails

Regular updating of AI models

For critical infrastructure, AI should not be allowed to operate without appropriate human and regulatory oversight.

8. Main Legal Principle

A useful approach is:

AI decision → cybersecurity check → human verification → legal review → action

Instead of:

AI decision → automatic action

This is particularly important where an AI error can affect electricity supply, financial rights, privacy or public safety.

9. Conclusion

Cybersecurity risks of AI systems are becoming increasingly important because AI is being connected with critical infrastructure, personal data and financial systems.

The main risks include hacking, data poisoning, adversarial attacks, privacy violations, manipulation of algorithms and incorrect automated decisions.

Indian constitutional law, the IT framework and data-protection law provide important safeguards. The principles developed in Puttaswamy, Shreya Singhal and Anuradha Bhasin are useful for ensuring that AI systems respect privacy, legality and proportionality.

Direct Supreme Court case law specifically dealing with cybersecurity attacks on AI systems is still limited. Therefore, existing constitutional and cyber-law cases are mainly used for broader legal principles.

Exam Line

“Cybersecurity regulation of AI systems seeks to prevent hacking, data manipulation, privacy violations and harmful automated decisions while ensuring that AI remains secure, transparent, accountable and subject to human and legal oversight.”