Uploaded July 2026 | Updated September 2026, 2 weeks ago
Use code TIS at incogni.com/tis to get an exclusive 60% off
The unstoppable computer virus may already have a name - Morris II. Unlike traditional malware, this generative AI worm doesn’t need you to click a suspicious link, open an attachment, or download a file. An AI assistant simply has to read the wrong email, document, or image.
This video explores how researchers created Morris II, the world’s first generative AI worm, and tested it against systems powered by ChatGPT, Gemini, and other AI models. Hidden prompt injections can potentially steal sensitive information, spread themselves through email, and turn infected AI assistants into part of the attack without the user realizing anything is wrong.
We also look back at the original Morris Worm of 1988, how it crippled thousands of computers, and why today’s AI agents could make the next generation of cyberattacks far more dangerous. As AI gains access to inboxes, files, calendars, passwords, and financial systems, a single malicious instruction could travel through entire networks almost automatically.
The biggest problem is that this may not be a software bug that can simply be patched. It could be a fundamental weakness in how AI understands language. When any sentence can become an instruction, language itself becomes malware.
00:36 - How Morris II Works
03:31 - The Original Morris Worm
05:41 - Why Morris II Is Different
06:36 - The Agentic AI Arms Race
08:00 - How the Infection Spreads
09:00 - The Near-Zero Cost of Attack
10:15 - The Flaw You Can’t Patch
11:12 - The Prompt Injection Problem
12:54 - Can Morris II Be Stopped?
14:14 - When Language Becomes a Weapon
Narrated by: Josh Risser
🔔 Don't forget to SUBSCRIBE! 🔔
SUGGEST A TOPIC: bit.ly/suggest-an-infographics-video
💬 Come chat with me: discord.gg/theinfoshow
🔖 MY SOCIAL PAGES
TikTok ► tiktok.com/@theinfographicsshow
Facebook ► facebook.com/TheInfographicsShow
📝 SOURCES: pastebin.com/UJDJCvDm
All videos are based on publicly available information unless otherwise noted.
Use code TIS at incogni.com/tis to get an exclusive 60% off
The unstoppable computer virus may already have a name - Morris II. Unlike traditional malware, this generative AI worm doesn’t need you to click a suspicious link, open an attachment, or download a file. An AI assistant simply has to read the wrong email, document, or image.
This video explores how researchers created Morris II, the world’s first generative AI worm, and tested it against systems powered by ChatGPT, Gemini, and other AI models. Hidden prompt injections can potentially steal sensitive information, spread themselves through email, and turn infected AI assistants into part of the attack without the user realizing anything is wrong.
We also look back at the original Morris Worm of 1988, how it crippled thousands of computers, and why today’s AI agents could make the next generation of cyberattacks far more dangerous. As AI gains access to inboxes, files, calendars, passwords, and financial systems, a single malicious instruction could travel through entire networks almost automatically.
The biggest problem is that this may not be a software bug that can simply be patched. It could be a fundamental weakness in how AI understands language. When any sentence can become an instruction, language itself becomes malware.
00:36 - How Morris II Works
03:31 - The Original Morris Worm
05:41 - Why Morris II Is Different
06:36 - The Agentic AI Arms Race
08:00 - How the Infection Spreads
09:00 - The Near-Zero Cost of Attack
10:15 - The Flaw You Can’t Patch
11:12 - The Prompt Injection Problem
12:54 - Can Morris II Be Stopped?
14:14 - When Language Becomes a Weapon
Narrated by: Josh Risser
🔔 Don't forget to SUBSCRIBE! 🔔
SUGGEST A TOPIC: bit.ly/suggest-an-infographics-video
💬 Come chat with me: discord.gg/theinfoshow
🔖 MY SOCIAL PAGES
TikTok ► tiktok.com/@theinfographicsshow
Facebook ► facebook.com/TheInfographicsShow
📝 SOURCES: pastebin.com/UJDJCvDm
All videos are based on publicly available information unless otherwise noted.
