Google Announces AI System Gemini After Turmoil at Rival OpenAI
Google unveiled its powerful new artificial intelligence system, Gemini, which surpasses any current AI technology in the market, including OpenAI’s ChatGPT. Despite spending most of the year on its development, Google stated that Gemini will not be widely available until early next year due to the need for extensive safety testing. The introduction of Gemini is Google’s latest attempt to showcase its AI capabilities following the disruption caused by ChatGPT’s launch last year.
Gemini’s release will serve as a test for Google’s advancements in AI and its commitment to faster product development. At the same time, rival OpenAI is currently undergoing a period of turmoil that has shaken the AI community, highlighting the unsettled hierarchy within the industry.
To ensure comprehensive testing, Google plans to involve external customers in evaluating the most advanced version of Gemini before its wider release. According to Demis Hassabis, CEO of Google DeepMind, Gemini represents the company’s most ambitious science and engineering project to date.
Under the Gemini umbrella, Google intends to offer a variety of AI programs to customers. Notably, the software boasts the ability to process various media formats such as audio and video, catering to the increasing demands of users relying on chatbots for multiple purposes.
Additionally, Google’s algorithms powered by Gemini will enhance products like Bard, a rival to ChatGPT, and enable mobile-phone features that operate independently of network connections.
Google claims that its most advanced version of Gemini, the Ultra version, outperforms OpenAI’s GPT-4 on industry benchmarks. Following testing with a select group of customers, the widely available software development kit for Gemini Ultra is expected to launch early next year. Google also plans to release a new chatbot named Bard Advanced, powered by Gemini Ultra.
To cater to a range of tasks, Google will release a less sophisticated version of Gemini, named Gemini Pro. This version will be accompanied by a related Bard-powered algorithm, surpassing competitors like the free version of ChatGPT in quality, as stated by Google.
At a developer conference earlier this year, Google CEO Sundar Pichai unveiled the Gemini project. Sergey Brin, Google’s co-founder, who has been less involved in recent years, dedicated substantial time to the Gemini project and collaborated with researchers in the newly constructed Charleston East building in Mountain View, California.
The appointment of Demis Hassabis in April as the head of Google DeepMind aimed to merge previously separate teams and optimize expensive computing resources necessary for developing powerful AI algorithms.
Google leveraged a wide range of internet data, including content from its YouTube platform, to construct Gemini. The company relied exclusively on its specialized AI chips, known as tensor processing units, to develop and refine the system.
During a press demonstration, Google executives showcased Gemini’s capabilities by examining physics homework problems, detecting issues in prewritten solutions, and generating correct answers.
The announcement of Gemini follows OpenAI’s recent restructuring, reinstating Sam Altman as CEO and appointing new board members. OpenAI’s GPT-4, released in March, is regarded as the most powerful AI system currently available, utilized by major companies such as Morgan Stanley and Salesforce. However, OpenAI’s Bard, which garners approximately 220 million monthly visitors worldwide, trails behind ChatGPT in terms of user base.
Google’s experimentation with Gemini extends to its search engine, exploring an AI-powered response format that provides text passages instead of conventional web links. While this feature is currently available to a limited number of users, Google intends to gather more feedback before a widespread release.
Google’s shares have risen by 48% this year, surpassing the broader Nasdaq Composite Index but lagging behind some of its tech peers such as Microsoft, which is OpenAI’s primary supporter.
Gemini’s ability to process various types of information, including text, code, audio, images, and video, surpasses existing AI systems. Google’s commitment to its own AI infrastructure, along with Gemini’s advanced capabilities, strengthens its position in the ever-evolving field of artificial intelligence.
