CERT-In’s New Advisory Unveils Hidden Cyber Threats

The advisory highlights critical vulnerabilities in AI models, outlines multiple attack vectors that malicious actors can exploit to compromise AI security and provides essential guidelines to mitigate threats. 

The Indian Computer Emergency Response Team (CERT-In) has issued a new cybersecurity advisory, CIAD-2025-0013, highlighting the growing risks posed by generative artificial intelligence (AI). The advisory raises concerns over evolving cyber threats, emphasizing the need for proactive cyber threat detection and mitigation strategies.

Key Cybersecurity Threats Identified

CERT-In’s report warns organizations and individuals about various new cybersecurity threats that cybercriminals are exploiting through AI-based attacks. These include:

● Adversarial Attacks: Hackers manipulate AI models by introducing deceptive inputs, leading to incorrect or harmful outputs.

● Model Inversion & Model Stealing: Threat actors attempt to extract sensitive data from AI models or replicate them for malicious purposes.

● Hallucination Exploitation: AI-generated misinformation can be leveraged to deceive users or spread false narratives.

● Backdoor Attacks: Cybercriminals implant hidden vulnerabilities in AI systems to gain unauthorized access.

● Prompt Injection: Attackers manipulate AI-generated responses by feeding deceptive commands, leading to unintended disclosures or actions.

CERT-In’s Cybersecurity Recommendations

To combat these cybersecurity threats, CERT-In advises organizations to enhance their security frameworks. CERT-In recommends robust AI model testing to identify and address vulnerabilities, preventing hacking threats and unauthorized manipulation. Implementing multi-layered security measures, including strong access controls, authentication mechanisms, and encryption, is crucial to safeguarding AI systems. 

Organizations should also establish advanced cyber threat detection and monitoring systems to identify and mitigate potential risks proactively. Additionally, developing a structured incident response plan is essential to ensure swift action against cyberattacks, minimizing potential damage and enhancing overall cybersecurity resilience.

As AI adoption accelerates across industries, the need for stringent cybersecurity measures has become paramount. CERT-In’s cyber alerts serve as a timely reminder that organizations must stay vigilant and adopt best practices to mitigate malware threats and prevent AI-driven cyberattacks.

By staying ahead of emerging hacking threats, businesses and government agencies can ensure a more secure digital environment in the evolving AI landscape.

Also Read: WEF’s report highlights CERT-In’s cybersecurity initiatives