On Wednesday, Google launched its new artificial intelligence model, which is considered its largest and most capable.
The language model, called Gemini, will feature three different sizes: Gemini Ultra, its most significant, most capable category; Gemini Pro, which ranges over a wide variety of tasks; and Gemini Nano, which it will use for specific tasks and mobile devices, CNBC reports.
“Gemini is the result of large-scale collaborative efforts by teams across Google, including our colleagues at Google Research,” CEO Sundar Pichai wrote in a blog post-Wednesday. “It was built from the ground up to be multimodal, which means it can generalize and seamlessly understand, operate across, and combine different types of information including text, code, audio, image, and video.”
Currently, Google plans to license Gemini to customers through Google Cloud to use in their personal applications. On Dec. 13, Gemini Pro will be accessible to developers or enterprise customers via the Gemini API in Google AI Studio or Google Cloud Vertex AI.
With the increased presence of AI-powered tools like ChatGPT, researchers warn that bots are experiencing “hallucinations.”
According to Jevin West, a professor at the University of Washington and co-founder of its Center for an Informed Public, a hallucination occurs when an AI model “starts to make up stuff — stuff that is not in line with reality.”
“But it does it with pure confidence,” West continued, “and it does it with the same confidence that it would if you asked a very simple question like, ‘What’s the capital of the United States?’”
West further explained that this means it would be difficult for users to distinguish whether what is true or not when asking a chatbot something they do not know the answer to, CNN Business.
Earlier this year, Google made headlines after its demo to Bird, a highly anticipated competitor to ChatGPT, provided an incorrect response to a question concerning new discoveries made by the James Webb Telescope. At the time, a Google spokesperson explained that the incident “highlights the importance of a rigorous testing process” and said the company was laboring to “make sure Bard’s responses meet a high bar for quality, safety, and groundedness in real-world information.”)
Suresh Venkatasubramanian, a professor at Brown University who helped co-author the White House’s Blueprint for an AI Bill of Rights, explained that fixing AI hallucinations remains a “point of active research” complicated to resolve.
“These models are so complex and so intricate,” Venkatasubramanian said; therefore, “they’re also very fragile.”
He added that even minor input changes can have “changes in the output that are quite dramatic.”
Other AI models launched earlier include the Sam Altman-led GPT-4, while OpenAI launched ChatGPT on GPT-3.5 last year.
Milton Quintanilla is a freelance writer and content creator. He is a contributing writer for Christian Headlines and the host of the For Your Soul Podcast, a podcast devoted to sound doctrine and biblical truth. He holds a Masters of Divinity from Alliance Theological Seminary.