Campus Technology

The White House has announced it is moving forward with a voluntary framework for testing the cybersecurity capabilities and risks of advanced artificial intelligence models, though the standards guiding those evaluations remain classified. Representatives from OpenAI, Anthropic, Google, Meta, Nvidia and other leading AI companies recently met with administration officials to discuss the framework.

For years, security researchers warned that attackers would use AI to improve existing cybercrime techniques. But according to Check Point Research's recent "AI Security Report 2026," the threat landscape has entered a new phase: AI is no longer just helping attackers; it is beginning to operate inside real-world attacks. The report describes a shift from AI as a force multiplier to AI as an active component of cyber operations.

Cyber attackers are increasingly turning the technologies and relationships organizations trust most — including cloud identities, AI tools, software dependencies and legitimate authentication processes — into paths for intrusion, according to CrowdStrike's latest threat-hunting research.

A Q&A with Eli Blouin and Mark Frydenberg Eli Blouin, currently a data analytics and marketing major at Bentley University, and Mark Frydenberg, Distinguished Lecturer of Computer Information Systems at Bentley ponder the day-by-day maturation of technologies that help us work and learn. Mark is the director of Bentley's CIS Sandbox, a technology social learning space, where Eli works as a student tutor. Increasingly, advanced technologies play a role that especially impacts student learning.

Artificial intelligence is raising the stakes of cyber conflict as attackers use the technology to accelerate reconnaissance, uncover vulnerabilities, and launch attacks faster than many security teams can respond. That widening speed gap is pushing organizations toward AI-powered systems capable of taking a more active role in cyber defense, according to Kai's 2026 State of Autonomous Defense Report.

AI security is entering a new stage as organizations look beyond protecting models from prompt injection, unauthorized access, and data exposure. Separate announcements from Microsoft and Nvidia recently outlined how AI could become an active part of cyber defense, with autonomous systems helping organizations detect threats, evaluate risks, and respond at machine speed.

After years of experimenting with generative AI, launching pilots, and testing copilots, enterprise organizations are now facing a more difficult challenge: turning AI into reliable, scalable business systems. According to Gartner's recent "Hype Cycle for Enterprise Architecture, 2026" report, enterprises are moving from AI experimentation toward a more industrialized approach to AI delivery.

NVIDIA and 36 other technology, cloud, cybersecurity, and enterprise software organizations have launched an industry effort to develop and share open technologies for protecting software and AI agents. The Open Secure AI Alliance initiative covers the infrastructure surrounding agents — including identity, permissions, isolation, harnesses, guardrails, logs, and evaluation systems — rather than focusing only on the security of AI models. The cloud angle is explicit in the alliance announcement.

Security researchers at Group-IB have uncovered a Windows malware strain that turns Microsoft 365 calendars into covert channels for receiving commands and stealing files from targeted organizations. The malware component, dubbed HOLLOWGRAPH, abuses the Microsoft Graph API and a compromised Microsoft 365 account to conceal command-and-control activity within legitimate cloud communications.

AI agents have quickly moved into mainstream enterprise use, but the content infrastructure needed to support them has struggled to keep up, according to a new report from cloud content management company Box. The survey found that 83% of organizations are running AI agents, yet only 36% of those using or experimenting with the technology have connected agents to trusted internal content across multiple use cases.