OpenAI is reportedly working on its next groundbreaking product: a ChatGPT that can use devices. The potential development of an autonomous AI assistant system capable of taking control of a user’s device to perform tasks has been revealed, according to a source familiar with the matter cited by The Information.
While OpenAI has not yet provided any official comment or clarification on the matter, it is logical to assume that action agents would be the next step beyond generative AI systems like ChatGPT. Currently, generative AI systems generate human-like media such as text, images, audio, and video. However, in order to make these models perform real-world actions like operating robots, developers must integrate them with external applications that adapt the AI’s output into programmable executables.
Compared to the technology behind smart assistants like Google’s Gemini or Amazon’s AI products, the underlying technology of ChatGPT is more robust. Thus, a virtual assistant developed with large language model technology, similar to ChatGPT, would have a greater potential for autonomous action.
The reported autonomous action agents developed by OpenAI are said to be capable of operating users’ devices to perform specific tasks. For example, a user could ask the AI to copy data from one platform to another. Essentially, any physical task a human can perform, such as swiping, tapping, clicking, double-clicking, typing, or even solving CAPTCHA puzzles, could be executed by an AI system with the necessary device privileges.
However, the road to autonomous assistance systems is fraught with privacy and security challenges. Current generative AI systems require connectivity to large cloud compute centers, and while some AI functions can run on laptops and smartphones, it is unlikely that an AI action agent with autonomy would rely solely on an onboard AI chip. This raises concerns about privacy, as well as the security threat of granting a corporate AI system unrestricted access to private information. The realization of an autonomous action agent could potentially introduce a critical new cyber threat with global implications.
In conclusion, OpenAI’s reported development of an autonomous AI assistant system capable of using devices marks a significant leap forward. Although there are concerns surrounding privacy and security, this technology has the potential to revolutionize the way we interact with AI systems and could lead to a new era of virtual assistants.
