Should I set up a personal AI agent to assist with my daily tasks?
—Seek help
As a general rule, I think relying on any form of automation in your daily life is dangerous when taken to extremes, and potentially alienating even when used at the right level. moderate, especially regarding personal interactions. ONE I have an agent Organize my task list and gather online links for further reading? Great. An AI agent automatically texting my parents every week with quick life updates? Terrible.
However, the strongest argument against incorporating more general AI tools into your daily routine remains environmental impact These models continue during training and produce output. With all that in mind, I did some digging WIRED archivepublished in the glorious dawn of this mess we call the internet, to find more historical context for your question. After a bit of searching, I came back and believe that you probably already use an AI agent on a daily basis.
The idea of AI agents, or God-forbidden “Agent AI,” is the buzzword these days for every tech leader trying to promote their recent investments. But the concept of one automatic assistant dedicated to completing software tasks is not a new idea. A lot of the discussion around “software agencies” in the 1990s mirrors the current conversation in Silicon Valley, where leaders at technology companies now promise a large influx of AI-powered agents are trained to do online tasks on our behalf.
“I see a problem with people questioning who is responsible for an agent's actions,” one person wrote. WIRED interview with MIT professor Pattie Maes, originally published in 1995. “Especially problems like agents taking up too much time on the machine or buying something on your behalf that you don't want. The agents will raise many interesting issues, but I believe we will not be able to live without them.”
I called Maes in early January to get her perspective on how AI agents have changed over the years. She remains as optimistic as ever about the potential for personal automation, but she believes that “extremely naive” engineers do not spend enough time addressing the complexities of human-to-human interactions. computer. In fact, she said, their recklessness could cause another AI winter.
“Right now, the way these systems are built is being optimized from an engineering standpoint, an engineering standpoint,” she said. “However, they are completely unoptimized for human-designed problems.” She focuses on how AI agents still work easy to be fooled or use misleading assumptions, despite improvements in the underlying models. And misplaced trust will cause users to trust answers generated by AI tools when they should not.
To better understand other potential pitfalls for personal AI agents, let's divide the vague terminology into two distinct categories: those that feed you and those that represent you.
Feeding agents are algorithms with data about your habits and tastes that search through streams of information to find what's relevant to you. Sounds familiar, right? Any social media a recommendation engine that fills my timeline with relevant posts, or an ad tracker that keeps showing me those marshmallows for the thousandth time Instagram can be considered a personal AI agent. In another example from a 90s interview, Maes mentioned a newsgathering staff who was fine-tuned to bring in the articles she wanted. That sounds similar to my Google News landing page.