IT Nightmare: AI Hallucinations and RMM Backdoors

Sysadmins are fed up with AI-generated nonsense and the quiet threat of legitimate remote tools used by hackers.
Icelandic

IT Nightmare: AI Hallucinations and RMM Backdoors · Avonetics

🎧 Podcast episode
Coming soon — Guðmundur & Hildur are taking this one into the studio.
Sponsored—As an Amazon Associate, Avonetics earns from qualifying purchases. Sony WH-1000XM5 headphones Whether you work in a noisy office or travel often, Sony WH-1000XM5 headphones deliver outstanding noise cancellation and exceptional comfort. They are ideal for anyone who wants to stay focused or enjoy rich, detailed audio anywhere. Check out the link in the episode description. Learn more →

The daily battle of sysadmins is no longer just about fixing broken computers. It is about dealing with AI-generated nonsense and the quiet threat of legitimate remote tools used by hackers.

One IT manager sat on the phone for two hours, troubleshooting a problem for a director who refused to let him remote into her computer. Instead, she sent him screenshots of ChatGPT responses. 'This is what CoPilot said,' she insisted. The manager was fed up. He had seen this before. Users sending AI-generated fixes that were often wrong.

Read nextÓgnaröld um helgina: Tölvuþrjótar fengu guðs-ham yfir kerfum þúsunda fyrirtækja

One user asked a new plant manager to turn on the 'Packaging and Transport' module in the ERP system. The manager told him there was no such module. She assured him there was. It turned out she just wanted to see the shipping screens. ChatGPT had told her they should be called 'Packaging and Transport.'

The issue is not just about AI. It is about understanding your environment. One IT pro said that AI responses from management were a definite factor in him retiring earlier than planned. Their AI slop as 'answers and suggestions' in email threads cost him more time than if they had just stayed quiet.

Your brand, right here.Reach story-obsessed listeners in 45+ languages → advertise on Avonetics

But there is more to it. Microsoft recently published a write-up about attackers abusing legitimate MSP360 RMM software to deploy ScreenConnect as a second remote access tool. The question is: how do you tell the difference between 'expected RMM' and 'someone just installed another legitimate one'?

The debate is ongoing. Some argue that AI is a waste of time. Others argue that it can be useful if you understand your system. The verdict is clear: AI is a tool, not a replacement for expertise.

Our hosts dig into this story on the podcast.

0:00
0:00
Link copied ✓
↑