AI agents are here. How much should we allow them to do?

Should I set up a personal AI agent to help with my daily tasks?

– Find help

As a general rule, I think that relying on any kind of automation in your daily life is dangerous when taken to the extreme and will probably alienate you even when used in moderation, especially regarding personal interactions. that Amnesty International agent Who organizes my to-do list and collects online links for further reading? Beautiful. An AI agent that automatically messages my parents every week with a quick life update? shocking.

However, the strongest argument for not involving the most productive AI tools in your daily routine is: Environmental impact These models continue to run while training and generating output. With all that in mind, I dug in WIRED archivepublished during the glorious dawn of this chaos we call the Internet, to find more historical context for your question. After researching a bit, I’m back convinced that you probably use AI agents every day.

The idea of ​​AI agents, or, God forbid, an “AI agent,” is the current buzzword for every technology leader trying to promote their latest investment. But the concept of A Automated assistant Assigning programming tasks to complete programming tasks is not a new idea at all. Much of the discourse around “software agents” in the 1990s mirrors the current conversation in Silicon Valley, where leaders at tech companies now promise a coming flood of creative agents using artificial intelligence and trained to do our online chores for us.

“One problem I see is that people will wonder who is responsible for a customer’s actions,” A. says. Wired interview With Professor Patty Mays of MIT, originally published in 1995. “Especially things like agents taking too much time on the machine or buying something you don’t want on your behalf. Agents will raise a lot of interesting issues, but I’m convinced we won’t be able to Live without them.

I called Mayes in early January to hear how her perspective on AI agents has changed over the years. She’s more optimistic than ever about the potential for personal automation, but is convinced that “very naive” engineers aren’t spending enough time tackling the complexities of human-computer interactions. In fact, she says their recklessness could lead to another AI winter.

“The way these systems are built, right now, has been improved from a technical standpoint, from an engineering standpoint,” she says. “But they are not at all optimized to deal with human design problems.” It focuses on how AI agents work Easily deceived Or resort to biased assumptions, despite improvements in the underlying models. Misplaced trust leads users to trust answers produced by AI tools when they shouldn’t.

To better understand other potential risks for personal AI agents, let’s break the nebulous term into two distinct categories: those that feed you and those that represent you.

Nutritional agents are algorithms containing data about your habits and tastes, which search through large pools of information to find what suits you. Sounds familiar, right? any Social media A recommendation engine filling a timeline with personalized posts or a constant ad tracker showing me those mushroom gummies for the umpteenth time. Instagram It can be considered an artificial intelligence personal agent. As another example from the 1990s interview, Mays mentioned a newsgathering agent who was fine-tuned to return the articles she wanted. This looks like my Google News landing page.