Google Gemini | Launch, Controversy, & Facts | Britannica
Google Gemini, a generative artificial intelligence (AI) model and chatbot created by the search engine company Google, utilizes large language models (LLMs) to interact with users and produce content.
Launch and Controversy
Originally named Bard, Google Gemini was introduced by Google CEO Sundar Pichai in February 2023 as an experimental conversational AI service powered by LaMDA (Language Model for Dialogue Applications). Despite the initial excitement, controversies quickly arose.
In June 2022, a Google engineer claimed that LaMDA was sentient, which Google denied. This led to internal turmoil within the company, resulting in the engineer being terminated.
In 
, Bard faced criticism when a promotional demo showcased an error where it provided inaccurate information in response to a query about the James Webb Space Telescope. This mistake caused Google's parent company, Alphabet, to suffer a $100 billion loss in market value.
Employee Testimonies
Following Bard's launch, Bloomberg News released internal messages from Google employees expressing concerns. Some employees described Bard as "worse than useless" and a "pathological liar," suggesting that Google had rushed its release without proper ethical considerations.
Subsequent Developments
In April 2023, Google consolidated its AI research groups, Brain and DeepMind, into Google DeepMind to enhance its AI advancements. By December 2023, Bard was rebranded as Gemini, with an updated version utilizing a new LLM that outperformed competitors like GPT-4 from OpenAI. Co-founder Sergey Brin returned to Google to contribute to the development of the new system.
Gemini Technology
Google categorized Gemini into three editions: Nano for accessibility features on Android devices, Pro for versatile performance, and Ultra for speed and quality. Gemini is multimodal, trained on text, images, and sounds, enabling it to combine various types of information effectively.
In response to user feedback, Google released improved versions of Gemini, including Gemini 1.5 Pro and Gemini 1.5 Flash, with enhanced performance and reduced software bulk.
Google debuts first Gemini generative AI model, with advanced 
...
Continued Expansion
Renamed from Bard to Gemini in early 2024, the platform expanded its language support beyond English to include Japanese and Korean. As of April 2024, the Gemini web app is accessible in over 40 languages, emphasizing its global reach.
Gemini Advanced and Criticisms
Google introduced Gemini Advanced for complex tasks such as coding and logical reasoning. Despite its capabilities, the platform faced criticism for generating inaccurate and offensive images, prompting Google to pause human image generation features and acknowledge the need for rigorous testing to avoid biases.
