Machine Learning

In the latest installment of our ongoing series where we identify and investigate security flaws in Azure Machine Learning (AML), we explore how cybercriminals could manage to covertly gain persistence in AML workspaces.

In our previous entry, we examined how credentials were being stored and logged in cleartext on compute instances (CIs) created in Azure Machine Learning (AML) service and the risks posed by the same. This article examines an information disclosure bug we found in one of the cloud agents used in the AML service and sheds light on the importance of threat modeling the agents’ features to uncover silent and hidden attack surfaces.

We discuss the present state of the malicious uses and abuses of AI and ML and the plausible future scenarios in which cybercriminals might abuse these technologies for ill gain.

We look into developments in end-to-end deep learning for cybersecurity and provide insights into its current and future effectiveness.

Trend Micro Cloud App Security detected and blocked 12.7 million high-risk threats that passed through the built-in security of cloud-based email services. We provide an in-depth report on the notable trends and stories from 2019 to help organizations prepare.