Researchers Create AI Web Agent That Allows People to Say and Get Things Done
The internet is such a vast and ever-evolving ecosystem comprising billions of websites and applications that have since become interwoven with our daily lives, shaping how we live and connect with people. However, it is not equally accessible to everyone.
The complexity involved with navigating the internet can be overwhelming for some people and particularly challenging for people with disabilities who may have difficulty with vision, hearing, mobility, or cognitive function. However, researchers at The Ohio State University are looking to resolve this problem with artificial intelligence.
AI Web Agents Turn Words into Action on the Web
According to reports, the researchers are using AI to develop web agents that can complete complex tasks on any website using simple language commands. More like online helpers, the AI web agent will help people complete open-ended tasks, including booking a flight, ordering some delivery, and several other functions just by saying it.
The study report for the AI web agent was initially released in December at the 37th Conference on Neural Information Processing Systems. In a recent interview, Yu Su, co-author of the study, explained:
“So with these web agents, what we want to achieve is that you can just go to any website and just say whatever you want to do, like book a flight, order some delivery, or even book a car knowledge test on DMV. Then the agent will understand what you said, understand this website, and just do it for you. That’s essentially the goal.”
A Better Internet Experience
The AI web agents leverage the power of both small and large language models to achieve things. According to the researchers, they can understand different layouts and functionalities of websites and can work similarly to how humans behave when browsing the internet, thanks to their ability to process and predict language.
While the AI agents will have a profound impact on improving the online experience for internet users, the researchers did note that there is also a “tremendous potential for harm.” They said autonomous agents could be exploited for potentially dangerous actions, such as misusing financial information or spreading misinformation.
“We should be extremely cautious about these factors and make a concerted effort to try to mitigate them,” said Su.
Land a High-Paying Web3 Job in 90 Days: The Ultimate Roadmap
