Human error and AI: The twin challenges of modern cybersecurity

In an era of LLMs and Generative AI, the problems around cybersecurity will only be compounded, underscoring the urgent need for action.

At a recent conference, the CIO of one of India’s leading banks made an honest confession, emphasizing the gravity of escalating cybersecurity concerns in the AI age. He pointed out that businesses often downplay these issues in public—a fact that should not be taken lightly. This understatement could be due to the fear of losing customer trust, attracting regulatory scrutiny, or facing unnecessary bottlenecks in their AI innovations.

Despite enterprises’ efforts to prioritize cybersecurity investments, human weakness remains one of the most persistent challenges. Numerous industry surveys over the past few years reveal that human errors are responsible for over 70% of organizational breaches. This stark reality highlights the need for technology leaders to focus on conducting comprehensive user training programs and enhancing capabilities for monitoring and controlling sensitive content.

In today’s world, data is considered as good as gold for organizations, and their success, innovation, and growth depend heavily on the security of their data protection strategies. However, executing their long-term and short-term goals remains challenging, especially when there is a lack of talent, resources, and huge disparate systems. 

In an era of LLMs and Generative AI, the problems around cybersecurity will only be compounded, underscoring the urgent need for action. Cybercriminals can inject malicious data into the data that an organization’s employees continuously feed into public and private LLMs to bias the model’s output, compromising the systems that rely on these models. If training data is not properly sanitized, information leakage and data poisoning can aggravate the situation. 

Read more: 7 Strategic Steps to Outsmart Modern Cybersecurity Threats

It is not that CIOs are unaware of these issues. However, as an industry, we must unite and collaboratively identify the best approaches to mitigate these threats and ensure they do not impede innovation. My colleague Praneeta, in her first cover story, engaged with many of you to grasp the gravity of the situation and the necessary steps to outsmart modern cybersecurity threats. While there is no foolproof strategy for complete security, it is crucial to focus on the most effective steps to remain vigilant and strengthen our approach to cybersecurity.

Share on