Cyber Security Headlines Week in Review: ChatGPT URL vulnerability, McDonald’s password problem, Perfekt Bluetooth blunder
14 snips
Jul 11, 2025 
Jim Bowie, VP and CISO at Tampa General Hospital, dives into the evolving threats in cybersecurity. He highlights the potential for ChatGPT to mislead users through phishing URLs. The discussion covers alarming vulnerabilities found in McDonald’s AI job application bot, raising questions about corporate security standards. Bowie also emphasizes the vital importance of basic cybersecurity hygiene, cautioning against neglect. With a light-hearted tone, they explore tech transparency issues and the risks of outdated medical devices and Bluetooth technologies.
AI Snips
Chapters
Transcript
Episode notes
LLMs and Phishing Risks
- Large language models like ChatGPT provide incorrect URLs about 34% of the time, creating phishing risks.
- Users should be wary of blindly trusting LLMs as reliable sources for sensitive information like URLs.
Managing AI Use Securely
- Instead of blanket blocking AI like ChatGPT, use proxies and firewalls to monitor and control AI web access.
- Hosting private instances of LLMs can help manage and filter their web queries for security.
Military vs. Infrastructure Cybersecurity Budgets
- The U.S. military cybersecurity funding is increasing focusing on advanced AI and offensive operations.
- Meanwhile, budget cuts to agencies like CISA risk undermining critical infrastructure defense and hygiene.