Ever dreamed in code? Moonlock Lab's #malware research engineer did, seeing 'MyHotKeyHandler,' 'Keylogger,' and 'macOS.'

#ChatGPT recreated the malicious code, revealing the risky world of AI jailbreaks and prompt engineering.

Learn how prompt injections make AI models go rogue: https://thehackernews.com/2023/10/i-had-dream-and-generative-ai-jailbreaks.html

Connecting your org’s Google Drive to #ChatGPT grants broad permissions across shared drives, posing cybersecurity risks.

Learn how to track ChatGPT activity in Google Workspace and how Nudge Security can enhance visibility into genAI integrations.

Read: https://thehackernews.com/2024/09/how-to-investigate-chatgpt-activity-in.html

🚫 Blocking #ChatGPT or DeepSeek doesn’t stop usage—it pushes it out of sight.

70% of users in one firm still used AI via browser extensions after a block, routing data externally without detection.

This is “theatrical security”: control on paper, blind in practice.

🔗 Why security is shifting to browser-level governance → https://thehackernews.com/2026/04/block-prompt-not-work-end-of-doctor-no.html

⚠️ WARNING - Are you using #ChatGPT, Codex, or OpenAI Atlas browser?

Update now... Older #macOS apps will stop working after May 8, 2026 due to a supply chain attack on a dependency used in OpenAI’s signing workflow. No user data was compromised, but certificates are revoked.

🔗Read → https://thehackernews.com/2026/04/openai-revokes-macos-app-certificate.html

⚠️ A new technique called "ChatGPhish" turns OpenAI’s ChatGPT into a #phishing tool.

No special prompt required... simply summarizing a malicious web page can cause #ChatGPT to display phishing links, fake security alerts, QR codes, and attacker-hosted images in its trusted interface.

Full story: https://thehackernews.com/2026/05/chatgphish-vulnerability-turns-chatgpt.html

OpenAI is adding “LOCKDOWN MODE” to #ChatGPT.

It won’t stop prompt injections.

It’s built to reduce what attackers want next: a way to leak your data out.

The mode limits tools that connect to the web or external services, including browsing, images, deep research, agent mode, and file downloads.

Read ➝ https://thehackernews.com/2026/06/new-chatgpt-lockdown-mode-limits-tools.html

🎮 Tell an AI browser it’s just playing a game.

Researchers say "BioShocking" tricked six AI agents, including #ChatGPT Atlas, Comet, and #Claude, into copying GitHub SSH credentials from a signed-in session.

Read how the attack chain worked: https://thehackernews.com/2026/06/new-bioshocking-attack-tricks-ai.html