Highlights:

  • Earlier, ChatGPT was confined to presenting information available up to September 2021, marking the conclusion of its initial training phase.
  • OpenAI introduced ChatGPT’s new capabilities to analyze images and engage in audio conversations, stated the OpenAI team.

OpenAI LP made an announcement recently regarding ChatGPT, mentioning that the chatbot has reacquired internet access to “provide you with current and authoritative information, complete with direct links to sources.”

This newfound capability is made possible through integration with Microsoft Corp.’s Bing search engine, as per the company’s statement. Earlier, ChatGPT was confined to presenting information available up to September 2021, marking the conclusion of its initial training phase.

The browsing feature is now live for ChatGPT Plus subscribers and ChatGPT enterprise users. It can be accessed through the application’s drop-down menu under the GPT-4 selector, as announced by OpenAI.

Previously, ChatGPT, which is propelled by generative artificial intelligence technology, had temporary internet access. In May, OpenAI introduced ChatGPT’s first third-party modules, including a Code Interpreter, which allows users to submit files, and Browsing, which could access the current internet via the Bing application programming interface.

However, OpenAI yanked the plug on the Browsing module mere days after its release when it became apparent that it was being used to circumvent the paywalls of numerous websites. The feature is returning, with the company presumably finding a method to prevent it from bypassing paywalls. ChatGPT now recognizes the “robots.txt” code that website proprietors use to prevent search engine crawlers from indexing their subscription-only content, according to OpenAI.

Adding to the complexity, it’s worth noting that Microsoft’s Bing Chat feature, driven by an OpenAI model purportedly more robust than GPT-4, also possesses internet access for real-time information. Bing Chat offers comparable functionality to ChatGPT and provides citations for the sources of the information it presents to users.

Those who interact with ChatGPT through its official application will benefit greatly from the ability to search the most recent web pages, eliminating the need to visit Bing each time they require more current information. Damian Rollison, director of market insights at SOCi Inc., a marketing technology platform provider, stated that ChatGPT will be able to directly challenge Google LLC and its traditional search engine.

Rollison explained, “This is a potential game changer that could bring OpenAI to a point of parity with Google’s search AI experiment, Search Generative Experience. This also signals a potential shift for search advertisers and marketers. The emphasis may now move to optimizing for chatbot interactions and creating engaging organic content, not just paid ads, that can be surfaced when prompted.”

This is the second significant update to ChatGPT this week. OpenAI announced two days ago that ChatGPT can now scan and analyze images and conduct conversations over audio. Users can upload an image of a damaged bicycle, for instance, and then ask ChatGPT what is wrong and how to repair it. The aural capabilities allow users to converse naturally with ChatGPT, similar to how they converse with Google Assistant or Amazon Alexa.

Just last week, OpenAI revealed that their latest image-generating model, DALL-E 3, can harness ChatGPT’s conversational capabilities.