OpenAI launches a new ChatGPT agent to aid users with more than just outputting impressive-sounding texts.
The advent of AI has had a ripple effect across the entire world, establishing itself as the steering wheel of the modern landscape.
Numerous tech giants and AI startups have invested their time, money, and infrastructure into gauging the maximum potential of this modern tech. But its potential remains unimaginable or even untouchable.
Meanwhile, with the knowledge we currently hold in our palms, humans are developing newer models, breaking down the limitations bit by bit. The comprehensible aspects of AI undergo several reiterations with the hope that it’ll set a new frontier for what is possible with this tech innovation.
And the interest remains surging.
OpenAI is the latest tech powerhouse to leverage this demand.
ChatGPT users will now be able to use a personal assistant, a new agent, to book flights and restaurant tables. These are the most basic tasks that the new ChatGPT agent will be able to perform.
This AI model will not just think, but also act. It’ll manage computer-based tasks, judge which software is perfect in a given situation, and autonomously complete tasks. The ChatGPT assistant will have control over your computer files, web browsers, and software, and combine them with the power of AI-driven research tools.
With the explosive demand for AI agents, there must be more they can aid with than just churning out flowery content. And be of real utility.
But all of it comes with added risks. Allowing AI agents to exercise control over computer systems would be fool’s work. It could easily fall prey to hidden prompts while it toggles across the websites it visits.
OpenAI’s intention behind it is to assist users with mundane tasks, such as reserving a table or making an itinerary. But they have also recognized the need to establish safeguards, given their potential. The tech company executed test trials and trained its model to reject similar suspicious requests for collecting a user’s data.
“You’re always in control. ChatGPT requests permission before taking actions of consequence, and you can easily interrupt, take over the browser, or stop tasks at any point,” said the company in response.