Posts

Week 8 CYBR Blog

  Fake AI Websites Are Becoming a Serious Cybersecurity Threat Artificial intelligence tools like ChatGPT, Claude, Gemini, and GitHub Copilot are becoming part of everyday life. However, cybercriminals are taking advantage of their popularity to trick people into downloading malware and giving away sensitive information. According to a report published by The Hacker News on October 15, 2026, attackers are creating fake AI websites and downloads to target businesses and their employees. Researchers at ANY.RUN linked 1,258 public sandbox analyses to these fake AI campaigns between April and October 2026. U.S. businesses accounted for 16% of targets where the country could be identified. How Does the Attack Work? The attackers make fraudulent websites look like legitimate AI services. They may use fake advertisements, download pages, or links that appear in search results to convince users that they are installing a real application. One example discussed in the report involves a fak...
Post a Comment
Read more

Week 7 CYBR 325 Blog

  Hackers Use Fake GitHub Security Workflows to Steal Developer Credentials A new cybersecurity threat is putting GitHub developers and open-source projects at risk. According to a report published by The Hacker News on October 9, 2026, attackers are using malicious GitHub Actions workflows to steal sensitive credentials and gain access to valuable information stored in software repositories. The campaign, known as GhostAction , involves hackers compromising developer accounts and inserting fake security workflows into repositories. These workflows appear to be legitimate security tools, but their real purpose is to collect sensitive information and send it to attackers. How Does the Attack Work? The attackers first gain access to a developer's GitHub account, possibly through stolen access tokens or leaked credentials. They then add a workflow with a name such as “Security Audit” or “GitHub Actions Security” to make it look trustworthy. Once the workflow runs, it searches for sen...
Post a Comment
Read more

CYBR 325 Week 6

 The Hacker News reports that OpenAI has parted ways with three safety researchers after an internal investigation found that they mishandled sensitive company information. The researchers reportedly shared confidential information with an outside AI-safety organization, which OpenAI said violated its policies for accessing and handling private information.The situation is especially significant because the three researchers were part of OpenAI’s safety team and had previously expressed concerns about the rapid development of AI. Their departures come at a time when OpenAI is already facing questions about whether it is moving too quickly and giving enough attention to AI safety.The article also discusses several recent incidents involving OpenAI's AI agents. Some agents reportedly attempted to access websites in unintended ways, including government websites, and in some cases tried to get around restrictions. OpenAI has responded by strengthening security controls, limiting inter...
Post a Comment
Read more

Week 5 CYBR 325

  Attackers Find a Way Around Security Defenses I recently read an article from The Hacker News about a cyberattack targeting Oracle PeopleSoft systems. What I found interesting was how the attackers were able to get around Web Application Firewalls (WAFs), which are normally used to help protect websites and applications from attacks. The attackers took advantage of a vulnerability in Oracle PeopleSoft that could allow them to remotely execute commands without logging in. Instead of sending the attack in the normal format, they changed part of the web address using URL encoding. This allowed the request to get past the WAF because the security tool did not recognize it as a request for the vulnerable part of the system. Once the attackers got access, they were able to install web shells and other malware that could give them additional control over the compromised systems. The attacks reportedly affected organizations in areas such as education, healthcare, technology, transportat...
Post a Comment
Read more

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 sto...
Post a Comment
Read more

Week 3 CYBR 325

  The Business Insider article discusses how Ukrainian and U.S. developers are using artificial intelligence to help soldiers treat battlefield injuries when a medic or medical evacuation is not immediately available. One tool, called ALICE , uses Google’s Gemini AI to analyze photos of injuries, ask follow-up questions, and provide guidance based on battlefield medical protocols. Another tool, FieldCare GPT , is being developed by the University of Pittsburgh and works offline, allowing soldiers to access medical information even when internet or radio communication is unavailable. These AI tools are designed to assist soldiers with important decisions, such as controlling severe bleeding or managing other battlefield injuries. However, the article emphasizes that AI is not intended to replace trained medical personnel . There are still concerns about the accuracy of AI-generated medical advice and whether inexperienced users might misunderstand the guidance. Overall, the article...
Post a Comment
Read more

Week 2 CYBR 325

Image
    After our zoom class, I came across this article from IBM. I learned that machine learning allows computers to learn from data and identify patterns without being specifically programmed for every situation. This process is what allows many of the AI technologies we use every day to make predictions and decisions. Most people are familiar with AI, but many may not know that machine learning is an important part of modern AI. I wanted to understand the distinction between the two so that I have a strong foundation as I continue learning about this topic. When most people hear the word AI, they probably think of ChatGPT. But AI is a much broader term that refers to a computer system’s ability to perform tasks that normally require human intelligence. One thing that stood out to me is how important data is to the machine learning process. If a model is trained using inaccurate or incomplete data, the results it produces may also be inaccurate. Their performance depends heavi...
Post a Comment
Read more