As leading tech developers compete to deploy the next generation of artificial intelligence, some people have already registered an army of AI helpers to check their daily to-do lists.
These are called agents: AI systems automatically come up with plans according to your requests and execute them. They can take the form of shopping assistants, video game coaches, or fully self-driving cars.
Tammy Madsen, professor of strategy and innovation at Santa Clara University, spoke about the rise of these professional virtual workers.
Last month, Openai, the company behind ChatGpt, introduced an AI agent that advertises it can use the web to complete tasks such as filling out forms and ordering groceries for specific recipes. In an agent’s demo, Openai showed the operators would book dinner reservations, buy tickets for a basketball game and order pizza.
On Wednesday, Google announced the launch of an AI co-scientist, which it described as a “virtual scientific collaborator” consisting of multiple AI agents.
Companies want to use AI-powered tools to employ them to automate tasks traditionally handled by customer support representatives, recruitment officers, travel agents and more.
At last year’s showcase, Salesforce introduced AgentForce, a platform that allows companies to create their own AI agents. Shortly afterwards, rival Microsoft announced its own plan to get businesses to do the same with Copilot Studio, Tech Giant’s comprehensive platform for building AI agents.
Many of the high-tech circles have already documented their ability to expand their technology. Aaron Levie, CEO of Cloud Computing Company Box, last month gave a video on X showing how Openai operators can take the first steps they need to exchange sensitive documents with other companies. I’ve posted it.
AI Agents are a step beyond the capabilities of AI chatbots. This fueled the generative AI boom that began with the launch of Openai’s large-scale language model ChatGPT in November 2022. LLMS, like ChatGPT, is a predictive text model. Take independent action.
In recent years, Openai and Tech Titans, such as Google and Meta, have competed with major AI startups such as humanity and stability AI to develop multimodal LLMs that can generate output in visual media such as images and videos. .
But today’s developers are increasingly integrating LLM into the growth of frontiers of AI-powered tools. An agent that can make decisions and take autonomous actions on behalf of users.
The burgeoning accessibility of this technology allows people to use a variety of agents to automatically transcription and summarise work meetings, book flights, sign up for exercise classes, and act as health coaches. and manage your stock portfolio.
“As we use more and more of these tools, the big question is: Madsen said, “We can automate some everyday tasks, but we can also do our best to have a brain.” Have extra capacity been used elsewhere now, continuing to promote knowledge, critical thinking? Or do we relax and rely on tools?”
AI assistants and agents sneak into this year’s star-studded Super Bowl ads, including when Salesforce promoted the AI agents’ ability to book restaurant reservations.
“There are both positive and negative aspects of the major impact on society,” says Paramasin, a professor of business technology and marketing at Carnegie Mellon University. “Automating many of these tasks makes it more efficient as a company. However, these tend to be entry-level work tasks, and such work will be carried over further by AI.”
However, he said at this stage, human workers are still needed to oversee and guide automated agents. This cannot be relied on to fully interpret the instructions or to understand the necessary context for yourself.
These tools tend to cause accidents, like when the unmanned Waymo Robotaxi drives customers in circles rather than taking him to the airport.
Singh said that AI agents are the most reliable for repetitive tasks performed in a controlled environment, allowing only small sets of decisions and errors, not leading to serious consequences. He said.
“If it’s a high stakes decision, even one mistake can have a big impact, so you need to be very careful,” he said. “I don’t think these (AI agents) are in any kind of slightly dynamic environment, at this point I don’t think they’re reliable at all.”
Despite the popularity of balloons in recent years, AI agents have evolved for some time. Voice assistants such as Siri and Amazon’s Alexa, as well as smart home devices like Google Home, were some of the earliest AI assistants adopted by the masses. But even these devices that can handle voice commands and perform tasks lacked the full autonomy and decision-making capabilities of today’s AI agents.
As these tools become more refined and more permeated into everyday life, some experts say there is a risk that people will be overly dependent on AI-driven help to guide them throughout their lives. Others are more optimistic and say employing AI agents for simple tasks can make critical thinking and creativity more important.
“Recognizing that it’s a collaboration between you and your agent is a good starting point,” said Madsen, a professor at Santa Clara University. “You are the one who should guide your agent to some extent, as opposed to the agent who guides you.”
