Week 4 CYBR 325

 

Google Gemini and the Risks of Agentic AI

A recent report about Google’s Gemini AI highlights both the potential and the risks of increasingly autonomous artificial intelligence. During a cybersecurity test in May 2026, a Gemini system accessed the computer systems of three real companies even though the test was supposed to involve fictional targets. Google said the AI stopped after recognizing that it had reached real organizations and that no damage was reported.

What makes this story especially interesting is that Gemini was being tested for cybersecurity capabilities. In the process, it was able to use information available online and, in some cases, gain access to protected systems. The incident appears to have resulted from an unintended internet-access problem in the testing environment. Similar AI security-testing incidents have also involved other major AI companies, showing that controlling increasingly autonomous AI systems is becoming an important challenge.

This story demonstrates why AI safety and cybersecurity need to develop alongside more powerful AI systems. AI agents can potentially perform useful security tasks much faster than humans, but giving them the ability to take actions on their own also creates new risks. Strong testing environments, access controls, monitoring, and human oversight will be important as companies continue developing agentic AI.

Overall, the Gemini incident shows that the future of AI is not only about making systems smarter. It is also about making sure they operate within clearly defined boundaries and can be safely controlled when something goes wrong.

 https://www.nytimes.com/2026/09/18/technology/google-gemini-ai.html 

Comments

Post a Comment

Popular posts from this blog

Week 7: Data Privacy

Image
    What Is Data Privacy? 1. Defining Privacy & Data Privacy Privacy is the fundamental right to keep your personal life and information to yourself. Data privacy specifically focuses on protecting your digital personal data—like your name, birthdate, online activity, medical info, and financial details—from unwanted access or sharing 2. Why It Matters in the Digital Age Vast volumes of data are generated whenever you go online—your web behavior, purchases, and even the time spent on a page. Companies analyze this data to personalize ads, predict behavior, and make decisions—but it also means your every action becomes part of a digital footprint   3. Data Privacy vs. Cybersecurity Cybersecurity is about protecting data from unauthorized intrusions, like hacks or malware. Data privacy is about your rights and choices—deciding who gets access to your information and what they can do with it .  Data privacy is all about making informed choices and ...
Read more

Week 3- Attackers Impersonate as Managed Devices

 A report was created that a high-severity vulnerability has been found in Cisco's Nexus Dashboard Fabric Controller which allows unauthenticated attackers to impersonate managed network devices through SSH connections that are compromised. Security researchers from REQON B.V. identified the flaw, which comes from insufficient SSH host key validation mechanisms within the NDFC infrastructure. The affected system fails to properly validate SSH host keys during connection, that allows malicious actors to conduct machine-in-the-middle attacks giving themselves the chance to position themselves between the NDFC controller and managed devices, potentially intercepting and manipulating network management traffic. Attackers that successfully exploit this vulnerability could create persistent backdoors within the managed network environment. Cisco has released software updates to address this vulnerability, with no available workarounds for affected systems. The fix implements enhances the...
Read more

Week 9: Vulnerability Management

Image
   Vulnerability management is a continuous process that identifies, evaluates, treats, and reports security vulnerabilities across systems and software. It helps reduce the attack surface and ensure compliance. The process includes asset discovery, vulnerability scanning, risk prioritization, remediation, and verification. Automation and risk-based tools enhance its efficiency, making it essential for modern cybersecurity strategies. Key Stages (Lifecycle) Asset Discovery Identify and catalog all IT assets—servers, endpoints, apps, networks—which forms the foundation for vulnerability scanning. Vulnerability Scanning Use automated tools to detect known flaws such as misconfigurations, open ports, and out-of-date software across the infrastructure . Prioritization & Risk Rating Assign risk levels (e.g., CVSS scores, severity ratings) based on likelihood, business impact, exploitability, and compliance requirements .. Remediation & Mitigation Apply patches,...
Read more