Google, upping the ante in the artificial intelligence arms race, finally launched its project Gemini on Wednesday, Dec. 6. The new model, which comes in three different sizes — ultra, pro and nano — is rolling out across Google's suite of products.
Part of this roll-out involves an enhancement to Google's AI-powered chatbot Bard, which will now use a version of Gemini Pro for "more advanced reasoning, planning, understanding and more."
Google called the enhancement the biggest Bard upgrade since the chatbot's launch.
Related: OpenAI CEO Sam Altman says that ChatGPT is not the way to superintelligence
Google said in a statement that the model will be available in English in more than 170 countries, with the company planning to expand it to support more languages and geographies at a later date.
Gemini, Google said, is additionally being brought to the Pixel 8 Pro phone, which the company said has been engineered to run Gemini Nano. Google said that the model will be rolling out across more products and services over the coming months, including Google Search, Ads and Chrome.
The company said it has already begun to experiment with Gemini in Search.
Gemini Ultra, according to the statement, is still in the midst of trust and safety checks, and will not be made available tot he general public until early next year, when Google plans to launch Bard Advanced, which will give users access to the more powerful model.
More Business of AI:
- The ethics of artificial antellligence: A path towards responsible AI
- Google targets Microsoft, ChatGPT with huge new product launch
- AI is a sustainability nightmare – but it doesn’t have to be
"This new era of models represents one of the biggest science and engineering efforts we’ve undertaken as a company," CEO Sundar Pichai wrote in a statement.
Google said that Gemini Ultra is the first AI model to "outperform human experts on" massive multitask language understanding, which uses a combination of 57 subjects to test real-world knowledge and problem-solving abilities. Gemini Ultra scored a 90% on the test.
The model, according to Google, was trained to be multi-modal from the start, meaning that it was designed to understand text, images and audio at the same time.
The tech giant said that it is working with a group of external partners to stress-test the models in an effort to ensure their safety; Gemini, Google said, has the "most comprehensive safety evaluations of any Google AI model to date, including for bias and toxicity."
Related: The ethics of artificial intelligence: A path toward responsible AI
The model, according to AI expert Gary Marcus, seems to be relatively on par with OpenAI's GPT-4.
"From a commercial standpoint GPT-4 is no longer unique," he said. "That’s a huge problem for OpenAI, especially post drama, when many customers are now seeking a backup plan."
Marcus, who is convinced that such large language models do not represent a pathway to a generally intelligent AI model, said: "The fact that Google, with all its resources, did NOT blow away GPT-4 could be telling."
The launch of Gemini comes in the midst of an ongoing debate about AI ethics and safety, in which many researchers and experts are concerned about active harms — algorithmic discrimination, bias, misinformation and political instability — and remain worried that such hype around the inevitability of superintelligent AI and existential threats will continue to overshadow these real-world harms when it comes to regulation.
Gemini's release, meanwhile, could bode well for (GOOGL) -) Google's stock.
Deepwater's Gene Munster said in October that Microsoft's cloud dominance over Google was only a temporary setback — once Google launches and integrates Gemini, he said, "the scale will tip back toward Google and we'll see a re-acceleration."
"What Google wants to do is take what you use Google for today and add onto that generative features," Munster said at the time. "As they add that, you're going to go to Google more over time and there's more opportunities for them to monetize search. Big picture: the Google story is intact."
Shares of Google, up roughly 48% for the year, began to rise following the announcement.
Contact Ian with tips via email, ian.krietzberg@thearenagroup.net, or Signal 732-804-1223.
Related: Marc Andreessen defends Silicon Valley in bold, tech-loving manifesto
Get exclusive access to portfolio managers’ stock picks and proven investing strategies with Real Money Pro. Get started now.