Renowned artificial intelligence research organization OpenAI has made significant strides in advancing AI capabilities. After a beta testing phase, OpenAI has formally incorporated an internet-browsing feature into its generative AI chatbot, ChatGPT. This eagerly anticipated feature comes roughly three weeks after its reintroduction in beta, following several months of hiatus.
OpenAI’s continuous innovation in artificial intelligence has set a new milestone, as it officially implements Bing-powered internet search in ChatGPT and unveils the beta version of DALL-E 3.
ChatGPT, which has made headlines over the past year, was primarily limited to data up to September 2021, inhibiting its real-time search engine capabilities. OpenAI began integrating internet services into ChatGPT in March, which was not without risks. Unlike a static training data set, the live web is not curated, potentially opening doors to misuse and algorithmic chaos.
In May, OpenAI commenced the rollout of web search via Bing, the search engine belonging to OpenAI’s corporate backer, Microsoft. Access was extended to the ChatGPT mobile app in late June. However, due to the unintended display of paywalled content, the feature was quickly withdrawn.
Fast-forward to late September, OpenAI reinstated the Browse with Bing feature, having refined how ChatGPT complies with instructions from content owners. ChatGPT now promises to respect a site owner’s Robots.txt file like traditional web crawlers. The improved Browse with Bing feature is now officially available to all Plus and Enterprise subscribers without any need to toggle their beta switch in settings.
DALL-E 3 Beta Release
In related news, OpenAI recently launched DALL-E 3 in beta, a month after unveiling the latest iteration of the text-to-image generator. The new version integrates with ChatGPT, easing image generation from text prompts. With DALL-E 3 directly embedded into ChatGPT, users can now receive photos as part of their text-based queries without switching between the two apps. DALL-E 3 is currently available in beta on the web and mobile, with users able to activate the feature by selecting “DALL-E 3 (Beta)” from the GPT-4 tab inside ChatGPT.
Expanding AI Capabilities
These developments are part of a broader expansion, steering ChatGPT from being a text-based generator to integrating audio and imagery. Last month, OpenAI equipped ChatGPT with speech capabilities, enabling users to converse verbally with the chatbot. This synergy of voice assistant technology with powerful large language models (LLMs) allows users to request inventive and on-the-spot narration of bedtime stories or similar tasks.
Additionally, ChatGPT now also allows users to search for answers using images. Users can upload a picture of an object and discover what it is or find similar items. These enhancements align with OpenAI’s broader objective to transform ChatGPT into a fully integrated, real-time, multimedia generative search engine.