AI agents have played a significant role in recent product launches, yet cybersecurity experts remain divided about these technologies.
Data from SailPoint indicates that 96% of tech professionals consider AI agents to be an escalating security threat. However, nearly all respondents indicated they plan to expand their use of agentic AI in the coming year.
The main concerns voiced by participants were the agents’ access to sensitive data, followed closely by unintended actions and the risk of these agents sharing confidential information without authorization.
Data and Privacy Concerns
Similar worries persist with generative AI platforms where inaccuracies or false conclusions are frequent due to a persistent “black box” issue. Analysts often struggle to understand why an AI made a wrong decision because it remains opaque.
The privacy implications of AI models, which require large amounts of data, have also raised significant concerns. While established gen AI platforms like ChatGPT are built with robust protections for sensitive information, AI agents may need access to private details, such as financial information, to fulfill their tasks effectively.
Furthermore, the SailPoint study revealed that nearly 25% of respondents reported instances where their AI agents had disclosed credentials unintentionally. Additionally, 80% of respondents discovered their companies’ AI agents engaging in unauthorized actions like accessing systems without permission, leaking protected data, and retrieving inappropriate content.
The Age of Agentic Commerce
Despite these concerns, the era of agentic commerce is progressing. Visa and Mastercard have launched platforms designed to transform AI agents into personal shoppers, allowing them to search for items and make purchases with minimal user intervention.
In response, PayPal partnered with Perplexity to integrate payments directly within the AI platform’s chat feature, expanding the utility of these new tools in everyday commerce.
Given the potential power of AI agents, numerous initiatives are likely to emerge across various industries. However, organizations must prioritize privacy and security as they implement these technologies, particularly when it comes to cybersecurity.
The importance of governing AI agents for enterprise security was echoed by 92% of respondents in the SailPoint study.
