Google has unveiled its ambitious AI model named Gemini, aiming to compete with OpenAI's GPT models. Gemini is touted as Google's "largest and most capable AI model" and is designed to be "multimodal," accepting inputs like text, images, audio, video, and code. Google plans to integrate Gemini into various products, including its search engine, Chrome web browser, and the Google Pixel 8 Pro, making it widely accessible. The company also highlighted its advances in cloud computing, claiming that its processors can train large AI models three times faster, potentially benefitting the broader AI industry. While Google acknowledges the risk of AI models providing misleading results, it has taken steps to improve factuality and is gradually releasing its most advanced version, Gemini Ultra, after safety evaluations.

‍

Google's launch of Gemini marks an effort to regain momentum in the generative AI field following the success of OpenAI's ChatGPT. The company's move also underscores its commitment to pushing generative AI across its ecosystem, catering to diverse user needs. By introducing a multimodal model and leveraging enhanced cloud computing capabilities, Google aims to stay competitive and ensure the responsible deployment of its advanced AI models. While Gemini shows promise in surpassing existing benchmarks, concerns about factuality and the potential for AI systems to generate incorrect information remain, prompting Google to implement additional techniques to enhance accuracy and safety measures.