OpenAI’s ChatGPT has become a worldwide sensation, causing a frenzy among top tech companies to develop their own ChatGPT-like AI. While others are trying to catch up, OpenAI seems focused on expanding its lead with new AI products that could automate even more tasks, potentially leading to job losses as a result.
As per a report from The Information, OpenAI is working on two kinds of AI agent software that would handle even more tasks typically performed by humans.
The first kind of AI agent might be employed to handle intricate tasks by essentially taking control of a customer’s device. For example, as per the report, the ChatGPT agent could move data from a document to a spreadsheet or complete expense reports and input them into accounting software.
These tasks usually involve actions like clicking, moving the cursor, typing, and other stuff humans do while working across different apps with data. Before this agent gets to work, it’ll need the user’s green light, as it’ll basically be running the show on their device.
Files related to the agent’s tasks might also need to be saved on the user’s device, and OpenAI might need permission from users to train the AI bot using personal data. The report also notes that OpenAI has trained its computer-using agent models on examples of humans using computers, including how we tackle various document formats like charts and PDFs.
The second type of AI agent is said to tackle tasks on the web, like collecting public data about a bunch of companies, planning trips, or even booking flight tickets.
We don’t have a clear timeline on when OpenAI will drop these agent products, but word has it they’ve been working on them for over a year now. With developments like these, AI is poised to take over roles that traditionally need human input and oversight because of how complex and varied the tasks are.