ChatGPT Gains Real‑Time Browsing and Multimodal Features – What It Means for AI

OpenAI officially announced that ChatGPT’s real‑time internet browsing feature is now generally available after a three‑week beta, ending a period of uncertainty.

When ChatGPT was first released, its knowledge was limited to data up to September 2021, and it would respond to requests for newer information with statements like “I don’t have the ability to browse the internet after September 2021.”

One of OpenAI’s goals is to extend ChatGPT’s knowledge and search capabilities to a multimodal, real‑time response system.

In March, the company piloted a web‑browsing plugin for a limited set of users. In May, OpenAI introduced a Bing‑based browsing feature, but after users discovered ways to bypass paywalls in June, the feature was paused. After allowing website owners to opt out of crawling, the beta was finally closed last Tuesday.

ChatGPT can now also generate images using OpenAI’s DALL‑E 3 model, though this capability remains in a testing phase for selected users.

OpenAI’s release notes state: “Ask ChatGPT what you want to see, from simple sentences to detailed paragraphs, and it will turn your ideas into highly precise images.”

Since its launch in November 2022, ChatGPT has become the fastest‑growing consumer app ever, reaching about 100 million monthly active users within two months. This rapid adoption sparked an AI “arms race,” prompting major tech firms to develop new features, applications, and experiences to differentiate their chatbots.

The two primary rivals in this race are OpenAI’s ChatGPT and Google’s Bard.

Earlier this year, Microsoft deepened its investment in OpenAI with an additional $10 billion, making it the largest AI investment in tech history. In April, the startup was valued at $27‑$29 billion after a $300 million equity raise backed by investors such as Sequoia Capital and Andreessen Horowitz.

Edited by Da Xiong