Plato Data Intelligence.
Vertical Search & Ai.

Bing web search market share barely moved by AI chat hype

Date:

Microsoft’s share of the global web search market has hardly changed since the arrival of Bing AI, aka Bing Chat aka Copilot, according to industry figures. We’re told Bing’s share has increased just 0.56 percentage points since it plugged OpenAI’s GPT-4 into its web search nearly a year ago.

The latest figures from StatCounter show that although Microsoft has attracted some fresh users to Bing after unleashing its conversational assistant, there aren’t that many of them. Google’s still the big daddy of internet search for the time being.

Microsoft launched its OpenAI-powered Bing chatbot in February 2023 when its global share of the search market across all platforms was 2.81 per cent at the time. Fast forward to December, and although it’s been increasing bit by bit every month, Bing reached just 3.37 percent, according to StatCounter.

The figures pale in comparison to rival Google, which had 93.37 percent of the global search market across all platforms in early 2023, although that slipped to 91.62 percent by December. On desktop alone, Bing climbed a little from 8.18 percent to 10.53 percent; Google fell from 85.64 percent to 81.71 percent. On mobile, Bing is still below one percent across the whole year, and Google has more than 95 percent of the global market.

It didn’t help that Microsoft limited the Bing chatbot to its Edge browser until the mega-corp started opening it up to things like Chrome and Safari around summertime. Edge is bumping along with just under five percent of the global browser market across all platforms, and about 12 percent on desktop. Microsoft also offers the Bing assistant via Android and iOS apps. We’ve asked StatCounter to confirm its market share figures for Bing include the chatbot.

Bing AI, today known as Copilot after a brief rebrand as Bing Chat, attempts to respond to your queries using natural language, and can offer summaries of pages and information, machine-generated content, and other stuff. When it was released, Google was criticized for being slow to deploy a rival chatty search assistant. Under CEO Sundar Pichai, the ad giant scrambled to catch up and rallied its AI engineers to create its competing oracle Bard, which was powered up publicly in March. It too tries to answer your questions and respond to requests in a conversational manner. Both Bing and Bard are known to hallucinate and make stuff up as all LLMs typically do.

Meanwhile, OpenAI’s GPT-4-powered ChatGPT became the fastest-growing app in history in 2023, partially helped by a $10 billion investment from Microsoft.

“We noticed an increase in usage by 10 times and that took us by surprise because if you think about it, DALL-E 2 was already quite good,” Jordi Ribas, Microsoft’s veep of search and AI, said today, sidestepping StatCounter’s numbers and referring to DALL-E 2, a popular image-generating bot built and launched by OpenAI prior to ChatGPT’s arrival.

“It really made a difference in the engagement and the users that came to our product.”

The bottom line is: all these hyped-up AI capabilities have barely increased Microsoft’s share of the global search market so far. Besides the Edge hurdle, it’s difficult to go up against Google considering it pays billions to be the default search engine on people’s devices. But money isn’t always everything and recent grumbling over the quality of the Chocolate Factory’s search results could spell trouble ahead for the super-corp.

Recent research showed that Google’s search quality, for one, seems to be declining due to the rising amount of SEO farms and affiliate link sites. The problem of low-quality content is only getting worse with generative AI churning out stuff at a large scale, and these weaknesses give competitors an opportunity to differentiate themselves and attract netizens away.

Perplexity AI, a startup that recently raised $73.6 million from Nvidia, Jeff Bezos, and others, is taking a shot at it. Self-described as an “answer engine,” it uses large language models to generate concise responses to users’ questions by extracting relevant information from websites. ®

spot_img

Latest Intelligence

spot_img

Chat with us

Hi there! How can I help you?