The race to secure AI is not just another chapter in the ongoing struggle between cybersecurity defenders and malicious actors
As artificial intelligence (AI) becomes increasingly integrated into daily life, concerns over its security are rising alongside the technology’s growing influence. AI is no longer a futuristic concept; it’s already reshaping industries, streamlining processes, and driving economic growth. In India alone, AI could contribute between USD 359 billion and USD 438 billion to the GDP by 2029-30, according to Michael Debabrata Patra, Deputy Governor of the Reserve Bank of India.
However, while much of the public debate around AI has centred on ethical questions, misinformation, and job displacement, a critical issue remains largely overlooked: the security of AI itself.
With AI systems managing tasks as diverse as medical diagnoses, financial transactions, and critical infrastructure, their vulnerability to cyberattacks poses a significant threat. A breach or malfunction in these systems could have catastrophic consequences, raising urgent questions about readiness and resilience.
Rising Awareness Of AI Security Risks
Recognising these challenges, India is taking steps to safeguard its AI landscape. The Ministry of Electronics and Information Technology recently held consultations to establish the India AI Safety Institute, aimed at ensuring the secure and ethical deployment of AI technologies across the country. This initiative is expected to strengthen domestic AI safety capabilities while fostering collaboration on a global scale.
In October, Chief of Defence Staff General Anil Chauhan launched the ‘Evaluating Trustworthy Artificial Intelligence (ETAI)’ Framework and Guidelines for the Armed Forces. He emphasised the importance of ensuring AI systems are not only effective but also resilient to attacks. His message applies equally to civilian applications as it does to defence systems.
Threat Vectors: How AI Can Be Attacked
There are two primary ways to compromise AI systems.
The first involves stealing data. AI models process vast amounts of sensitive information, ranging from personal health records to corporate secrets. Hackers can exploit these systems, tricking them into divulging secure information or compromising medical and financial databases.
The second threat lies in sabotaging the models themselves, leading to dangerous outcomes. Imagine an AI-powered car misinterpreting a “Stop” sign as “70 mph.” Such risks illustrate the critical need to protect AI from manipulation.
As AI adoption grows, so too does the list of potential attack scenarios. Abandoning the technology due to these risks, however, would be counterproductive, leaving organisations dependent on third-party solutions and unable to control a vital aspect of modern operations.
Strengthening AI Security: Key Steps
To reap the benefits of AI while managing its risks, experts recommend three key strategies.
First, organisations must choose their AI models wisely. Some AI types, such as large language models, are more vulnerable due to their reliance on vast datasets. In contrast, symbolic or hybrid models, which operate on explicit rules, are often more secure.
Second, proven defences such as digital watermarking, cryptographic solutions, and stress-testing tools like Thales’s “Battle Box” should be deployed. These measures can identify and address vulnerabilities before attackers exploit them.
Finally, organisational cybersecurity must be upgraded. AI is part of a broader information ecosystem, making it essential to strengthen traditional cybersecurity measures. Training employees to identify and prevent security breaches is critical, as human error remains a common weak point.
High-Stakes Battle
The race to secure AI is not just another chapter in the ongoing struggle between cybersecurity defenders and malicious actors. This time, the stakes are higher.
If AI security is not prioritised, there’s a real risk that the technology could be manipulated for harmful purposes. As AI becomes as integral to modern life as electricity, safeguarding its integrity is a responsibility that businesses, governments, and society must collectively embrace.
