SEATTLE WA- Google is the latest of big-tech companies to announce it is joining the race to produce the next-generation artificial intelligence (AI) tool, called an “agent.” It is envisioned to be a web browser-based tool that, with human direction, could tackle both mundane or complex chores.
One potential example: tell a search engine to “find tropical vacations” within a certain geography, dates, length of stay, favored flight times, and a not-to-exceed budget amount.
That typed command could send an agent out onto the Internet, to later return a list of suggested resorts at various price points. You may also be able to click and choose your preferred airline, model of rental car, and other options.
Big Tech is racing for a future in which artificial intelligence bots may be able to accomplish that and other kinds of shopping for you, according to Dr. Chirag Shah, a University of Washington Information School professor. Many “organizations now see this as the next phase” of AI, he says, “where systems can move beyond just generating information and can use that information for reasoning and taking action.”
Such autonomous computer programs “can function kind of like personal assistants,” Shah adds. In some ways, he notes, they’re already in use. Right now, some do everything from simply turning on room lights, to driving your car for you.
Google released its version of next-gen AI agents, called “Project Mariner,” on Dec. 11 (2024; Wednesday). Dozens of well- and lesser-known companies also are competing to be this next AI-driven “big thing.” Considering the progress to date, Shah believes “we will likely start having these kinds of capable agents soon.”
Worries About Personal Assistants
Not everyone will embrace such personal assistants, Shah acknowledges, despite their ability for users to eliminate some tasks they find tedious. A percentage of potential users may in fact like such tasks.
One worry is whether agents can be trusted. Building that trust will take time, he admits, and may likely start with small shopping steps.
What if AI makes a mistake? “There is a cost to making a mistake,” Shah agrees.
“Imagine this shopping scenario: you give the budget, you give the parameters and you get some outcome you’re not happy with.” Maybe you’re stuck with a choice, because an agent dealt with a company that doesn’t take returns or offer refunds. “You’re going to bear the cost, not the agent,” he points out.
Images by Ivan Vetushuko and YuliyaKirayonakBO, on Deposit Photos,
used by Travels With The Post under license
Foodie: At Barrister’s, Oven’s Hot and Producing Bagels