You’ve probably used AI to get information, but have you ever considered letting one run your life for you? AI “agents” promise to transform our daily routines by taking over mundane tasks like ordering food, shopping, and even answering emails. I decided to put this to the test to see if these digital assistants are truly ready for the job, or if it’s a recipe for disaster.
🤖 Autopilot or Incompetent Intern?
My experience was a mixed bag. When I asked an AI agent named Manus to reformat a presentation, it placed every single line of text in a separate box, making future edits a nightmare. Another agent, Operator, tried to submit an invoice for an article to New Scientist for an absurd £8001. It often felt less like having a super-intelligent assistant and more like managing a well-meaning but hopelessly confused intern.
🔒 The Trust and Privacy Dilemma
To truly embrace these tools, you have to hand over the keys to your digital kingdom—your passwords, your contacts, and even your credit card details. This raises significant security and privacy concerns. When I reluctantly typed my credit card info into a window controlled by Operator, I felt an enormous sense of unease. Researchers have found that agents can be tricked by malicious actors into clicking hidden links and giving away sensitive information, with no effective mechanism for user intervention.
🚀 A Glimpse of the Future?
Despite the frustrations, there were flashes of brilliance. The agents successfully compiled code and even managed to schedule an interview for this very story. Tech companies are betting big on this technology, but experts agree that we are still in the very early days. Fundamental breakthroughs are still needed before AI agents can become the reliable, autonomous assistants we’ve been promised. For now, it seems I’ll still be ordering my own prawn crackers.
—
Stokel-Walker, Chris. “AI at your service.” New Scientist, vol. 267, no. 3551, 12 July 2025, pp. 34-37.
More Topics
- Why Vacuum Energy is a Bigger Puzzle Than Dark Energy
- Have Scientists Found the First Known Human-Neanderthal Hybrid Child?
- Could Hackers Use Your Home Solar Panels to Disrupt the Power Grid?
- Is Your Nighttime Light Exposure Increasing Your Risk of Heart Disease?
- Are Humans Naturally Cooperative or Inherently Selfish?
- Is Artificial Intelligence About to Revolutionize Mathematics?
- Could a Single Drug Injection Protect Us From All Flu Strains?
