Developers of generative neural networks, such as Google and OpenAI, are confident that in the near future, users will delegate the processing of Internet requests and the execution of various tasks to chatbots or AI agents. Microsoft AI head Mustafa Suleyman believes that AI algorithms could replace web browsers, and some experts predict that such applications will eventually become similar to operating systems.
Image source: techspot.com
In a conversation with The Verge, Suleiman spoke about a future in which conversational generative AI interfaces could make regular web browsers obsolete. However, before this concept can be put into practice, the technology must overcome many obstacles. He expressed confidence in the further development of AI, but tempered expectations for artificial general intelligence (Artificial General Intelligence), which is also called strong AI that can solve problems like humans.
Suleiman believes that in its current form, the process of interacting with search engines is too cumbersome. In his opinion, it is much more convenient to make requests in the Copilot application on your iPhone. According to Suleiman, AI assistants with interfaces generated by neural networks will be able to replace traditional search engines and browsers within three to five years.
If developers redesign web portals to host AI agents, those agents will be able to complete requests by interacting with other AI systems. However, in this scenario, it is not entirely clear whether AI will be able to correctly find accurate answers to queries, and what will happen to advertising revenue when people stop clicking on websites. Suleiman played down those issues, saying Microsoft has made significant progress in minimizing the amount of AI hallucinations that are a major problem with AI agents.
Industry analyst Om Malik believes generative AI applications could go even further. In a recent post, he spoke about the limitations of traditional browsers in a world increasingly dominated by augmented reality, virtual reality, and artificial intelligence. Browsers haven’t changed much since their inception, they’re built around a “document-centric internet.” The emergence of generative neural networks and conversational systems challenges these foundations. AI has the potential to fragment web pages into interactive, personalized streams of information that defeat the original purpose of browsers.
Google is testing something similar as part of Project Jarvis, whose participants have created an AI agent capable of limited control of the mouse cursor and the user’s browser to automate the process of performing a number of tasks. Former Google employees are also working to create a working online operating system powered by AI.