Credit: Shutterstock
Microsoft made DALL-E generally available for its Bing AI assistant Tuesday, just a day before Google grafted Bard AI onto its Assistant app, as the battle over AI-driven search continues to heat up.
The stakes are high. In the ongoing Google antitrust trial, where the search giant is accused of maintaining an illegal monopoly on search via exclusive contracts with device makers like Apple, Microsoft CEO Stya Nadella testified about central role search plays in the tech industry. “In some sense, when you look at a company like us … there is nothing more attractive internet search,” Nadella said. “The organising layer of the web is internet search.”
With Google having captured 90% of the search market so far, Microsoft has backed DALL-E maker OpenAI as generative AI opens up new opportunities to attract search engine users.
DALL-E first rose to prominence in 2022, initially as a form of online entertainment. It’s an advanced image generator capable of taking a basic prompt — e.g. “Fenway Park but populated by aliens” — and creating an on-the-spot picture that tallies with what a user might expect to see. DALL-E version 3 replaces the pre-existing Bing Image Creator in Microsoft’s AI assistant.
Microsoft on Tuesday said that the latest version of DALL-E features several improvements, including better attention to relevance and an improved ability to follow natural-language prompts. It also boasts improved photorealism and coherence, as well as an additional level of stylisation and creativity.
Microsoft clearly views Bing and its AI-enabled features as a core consideration moving forward. During his testimony in the Google antitrust trial, Nadella said the GPT-powered version of Bing has pushed the company’s market share higher.
“GPT has led Bing to an all-time high,” Nadella said. “I’m glad we’re out there using a new technology.”
Will Google lock Microsoft out of AI search?
He expressed concerns, however, that Google’s continued dominance could lock his company, as well as other competitors, out of the AI-powered search market in the future. At issue is access to online content that can be used to train AI in order to make them more effective, which, according to Nadella, means that exclusive contracts for the use of that data make developing competition to Google unduly difficult.
Nadella’s fears are unlikely to be allayed by the news that Google will integrate Bard, its own in-house AI technology, into Google Assistant, which competes directly with Bing and its own AI assistant features. The search giant said that the idea is to dramatically expand the helpfulness of the Assistant functionality, making it more aware of context and giving it more robust functionality.
“For example, say you just took a photo of your cute puppy you’d like to post to social media,” according to a Google blog post. “Simply float the Assistant with Bard overlay on top of your photo and ask it to write a social post for you. Assistant with Bard will use the image as a visual cue, understand the context and help with what you need.”
The broader question — of whether these AI assistants are complementary, unrelated or rivals to traditional search engines — remains difficult to answer concretely. GPT-like AI doesn’t look at live information on the web to answer queries, it simply bases its answers on its corpus of training data. Hence, up-to-the-minute information won’t be provided and, as has been widely written about, it is likely to produce false results in many situations. Nevertheless, the way in which both Assistant with Bard and Bing AI have have been deployed demonstrates that search and AI are closely interrelated.
