Google Launches Gemini, Upping the Stakes in the Global AI Race
Google has unveiled its latest advancement in artificial intelligence (AI) with the introduction of project Gemini. This highly anticipated AI model is trained to behave in human-like ways, and its launch is expected to ignite further discussions on the potential benefits and risks associated with AI technology.
The rollout of project Gemini will occur in stages, with two less sophisticated versions called Nano and Pro being immediately integrated into Google’s AI-powered chatbot Bard and its Pixel 8 Pro smartphone. With Gemini’s incorporation, Google promises that Bard will become more intuitive and improve its performance in tasks involving planning. On the Pixel 8 Pro, Gemini will be capable of quickly summarizing device recordings and providing automatic messaging replies, starting initially with WhatsApp.
However, the most significant advances from Gemini are slated to arrive early next year with the introduction of the Ultra model. This model will be utilized to launch Bard Advanced, an enhanced version of the chatbot that will initially be available to a select test audience.
Although initially functioning exclusively in English worldwide, Google executives have assured that Gemini has the potential to expand into other languages in the future.
During a demonstration for a group of reporters, Google showcased the capabilities of Bard Advanced, which exhibited unprecedented multitasking abilities by simultaneously recognizing and comprehending presentations encompassing text, photos, and video.
While Google did not disclose the exact parameter count of Gemini during a virtual press conference, the company released a white paper outlining the model’s superior performance compared to GPT-4 in multiple-choice exams, grade-school math, and other benchmarks. However, Google also acknowledged the ongoing challenge of developing AI models with higher-level reasoning skills.
Computer scientists have highlighted some limitations in large language models like Gemini, which function by predicting the next word in a sentence and are susceptible to generating errors known as hallucinations.
Regarding Gemini’s factuality, Collins, a Google representative, stated, We made a ton of progress in what’s called factuality with Gemini. So Gemini is our best model in that regard. But it’s still, I would say, an unsolved research problem.
The launch of project Gemini by Google signifies a major advancement in the AI field, with potential applications ranging from improved chatbot capabilities to enhanced smartphone features. As Google continues to refine its AI models, experts recognize that certain challenges remain, particularly in the development of higher-level reasoning skills and ensuring factual accuracy. Nevertheless, the emergence of Gemini cements Google’s commitment to pushing the boundaries of AI and solidifies its position in the global race for AI supremacy.
