- Google delayed the launch of its new AI chatbot “Gemini” to next year.
- Launch was delayed by CEO Sundar Pichai
- It is due to issues with its prompt responses.
Google has reportedly delayed the launch of its highly anticipated generative AI chatbot and GPT-4 competitor called ‘Gemini’ until January. The news follows reports that Alphabet CEO Sundar Pichai was not convinced with the chatbot’s response to non-English queries.
The Silicon Valley behemoth teased Gemini at the Google I/O 2023, describing it then as a next-generation AI model with “impressive multimodal capabilities”. According to Google, Gemini can process multiple types of data and can understand and generate text, images, and other types of content, such as websites, based on a sketch or written description.
Google Delays Gemini Launch Due To Failure To Process Non-English Queries
Google initially planned to debut Gemini next week in a series of events held in California, New York, and Washington but has now quietly rescheduled it for early 2024. Two anonymous sources familiar with the unannounced event said that there were concerns with the AI’s struggles when it came to responding to non-English prompts and queries. This prompted Sundar Pichai to delay the launch.
Google aims to dethrone OpenAI’s GPT-4 and is targeting global language support for its generative AI model. In November, Pichai said that the company is focused on getting out “Gemini 1.0” as soon as possible. The Google chief added that the team is making sure that “it’s competitive” and “state of the art “.
Related Readings: 7 Best CySEC-Regulated Forex Brokers of 2023
Gemini Is A Multimodal AI That Is Capable Of Developing Websites From Sketch
At I/O, Google representatives said Gemini was showcasing impressive multimodal capabilities that were yet to be seen in prior models. Apart from understanding text and image contents, the company wants to make Gemini highly efficient at tool and API integrations, turning it into an attractive offering for third-party developers.
According to the company, Gemini is built to “enable future innovations, like memory and planning” and once fine-tuned and “rigorously tested for safety”, the AI model will be available in “various sizes and capabilities”.
While yet to be released for public use, Google touts Gemini as being capable of outperforming GPT-4 as it harnesses more computing power than its chief rival.
Talking to Business Insider, Sissie Hsiao, Google VP and GM of Bard and Google Assistant, said that she has seen Gemini do “some pretty amazing things”. She prompted the chatbot to draw three pictures of the steps on how to bake and ice a three-layer cake and it generated those images.
Hsiao said the pictures were completely “novel” and not sourced from the internet. She added that Gemini can speak in imagery with humans and not just via text like other models.
Google Expects Gemini To Enhance Its AI-Enhanced Products And Dethrone GPT-4
Google already has its own generative AI model called Bard, which has been integrated into its search engine, and other AI-enhanced products like the Google Assistant. The company expects Gemini to bring major improvements to its existing AI-related offerings.
Unfortunately for Google, ChatGPT, an AI model that has procured over $13 billion in investments from Microsoft, has so far enjoyed stronger consumer awareness and is dominant in the AI assistant space.
Therefore, it was logical for the company not to risk rushing Gemini out of the door given the significance of the product at hand. Analysts argue that Google’s multimodal AI could be a game changer for the industry once it’s launched.