Hackers are getting creative, and it’s no longer just about the robot.
As AI moves into robots, autonomous vehicles, and industrial systems, attackers are shifting their focus to the vulnerable underbelly of these complex technologies: the credentials, cloud services, and update channels that control them. ESET, a cybersecurity firm, has been at the forefront of this battle, blocking over **3,000** artificial intelligence (AI) skills it identified as malicious and flagging a further **25,000** as suspicious since March 2026.
Cloud Security Takes Center Stage
One of the key battlegrounds in this emerging threat landscape is cloud security. As more AI applications move to the cloud, attackers are exploiting vulnerabilities in cloud services to gain unauthorized access to these systems. This can have disastrous consequences, from crippling industrial processes to putting lives at risk in autonomous vehicles.
Malicious AI Skills on the Rise
ESET’s findings offer a stark reminder of the scale and speed of this threat. The firm has been tracking a surge in malicious AI skills, with many of these skills masquerading as legitimate tools for training and developing AI models. However, once deployed, these skills can be used to steal sensitive data, disrupt critical infrastructure, or even take control of entire systems.
What this means is that organizations deploying AI in their operations must rethink their security strategies to account for the cloud-based services that underpin these technologies. This requires a more nuanced approach to security, one that acknowledges the complex interactions between AI, cloud services, and update channels.
Security in the Age of Embodied AI
As we move into a future where AI is increasingly embodied in physical systems, the threat landscape is shifting to reflect this new reality. Embodied AI security, in other words, extends far beyond the robot itself to the networks, services, and credentials that control it. By understanding this shift and taking proactive steps to secure their cloud-based services, organizations can reduce the risk of catastrophic attacks and ensure the safe deployment of AI in their operations.
