https://damnsec.com/tags/ai/Recent content in AI on DAMNSECHugo -- 0.151.0en-usWed, 22 Apr 2026 00:00:00 +0000https://damnsec.com/blog/all-new-and-improved-cybersecurity-attack-surfaces/Wed, 22 Apr 2026 00:00:00 +0000https://damnsec.com/blog/all-new-and-improved-cybersecurity-attack-surfaces/<p>Since the generative AI boom, everyone has found a use case for it. Reviewing their writing, planning a study guide, sharing hard-coded API keys/company secrets critical to their organisation&rsquo;s infrastructure, and even using it as a financial adviser. What? The whole sharing company secrets thing? Pfft. Oh yeah, people do it all the time.</p> <p><a href="https://securityaffairs.com/144597/security/samsung-data-leak-chatgpt.html#:~:text=%E2%80%9CIn%20another%20case%2C%20an%20employee,Affairs%20as%20your%20favorite%20blog.">Samsung employees unwittingly leaked company secret data by using ChatGPT</a>. They shared information like meeting documents and source code with chatGPT. This led to 3 data leaks in just under a month via chatGPT. How about the <a href="https://thehackernews.com/2025/02/12000-api-keys-and-passwords-found-in.html">12,000+ API Keys and Passwords Found in Public Datasets Used for LLM Training</a>? There were 219 different types of secrets found here. AWS root keys, Slack webhooks, Mailchimp API keys, name it. A lot of these were live, too.</p>