Coming as an extension to GPT Chat, DALL-E adds to the Bing arsenal with which Microsoft is trying to dethrone Google Search.
DALL-E 2 can generate impressive images based on simple text descriptions. But the artificial intelligence-based service will likely be even more powerful after integration with Bing Search. For example, when Bing Chag sends you a query, you’ll get both detailed text answers, backed up with links to the sites where the information was cited, and automatically generated images. It doesn’t take much imagination to realise that a simple text search can turn into an extremely compelling presentation, supported by detailed and well-written text and suggestive images.
It remains to be seen whether the DALL-E vs ChatGPT duo will exacerbate fears about the dangers of artificial intelligence technologies as tools for disinformation and manipulation. Already, ChatGPT responses are notorious for the persuasive power they exert by delivering highly reasoned information, yet based on inaccurate, or downright malicious, sources. Add to that self-generated images to support the argument, and you have a recipe that would make even the most devoted propagandists affiliated with authoritarian regimes envious.
Based on an advanced version of OpenAI’s DALL-E model, Bing Image Creator allows users to create images describing the subject using their own words. To the exasperation of real artists, the Microsoft-developed AI draws its creations from a huge collection of images available “freely” on the Internet, from classic paintings like the Mona Lisa to the deeply intricate digital images artists have created with programs like Photoshop. The result: users can now generate original images of exceptional quality that cannot be attributed to a particular source of inspiration. In effect, DALL-E dilutes the importance of truly authentic creations delivered by real artists, in effect leaving them without the object of their work.
Bing Image Creator will be fully integrated into the Bing Chat experience. But initially users will only be able to access the less reality-bound Creative mode, with search modes centered on the accuracy of generated images to be added later. Whichever mode is selected, users will be able to create images by providing a simple text description, optionally specifying context and artistic preferences.
Microsoft Edge will also get a dedicated button for accessing the Bing Image Creator, making it the first web browser to omit an AI-based image generator.