SAN FRANCISCO — Google said it will soon make its own artificial intelligence chatbot available to the public and begin using the tech to generate answers in search results, firing back at accusations the company, long a leader in AI tech, has been slow to respond to competition from its rivals.
The search giant, which has invested huge amounts of money in AI research over the last decade, will make a chatbot called “Bard” publicly available in the “coming weeks,” according to a Monday blog post from Sundar Pichai, the chief executive. Google has been making a series of announcements on its plans for new AI tools and products in the wake of archrival Microsoft signing a multibillion-dollar deal with AI start-up OpenAI, which won spades of media and consumer attention after making its ChatGPT chatbot available to the public in November.
Google has been at the forefront of AI research for years, scooping up many of the field’s brightest scientists and using the tech to improve the quality of language translation, search results and a host of other technologies the company uses. But over the last six months, smaller companies like OpenAI have captured more attention — and venture capital investment — by making tools like AI image- and text-generators directly available to the public. That’s at odds with the Big Tech companies’ generally more cautious approaches, which have been shaped by earlier public relations disasters, such as chatbots that spouted racism and hate speech, or a Google project to build image recognition software for the military that spurred an employee revolt.
Now, Big Tech companies, especially Google, Microsoft and Facebook, are moving faster, causing fresh concerns among AI safety and ethics experts that the tech could be deployed too quickly before its consequences are fully understood.
“We’ll continue to be bold with innovation and responsible in our approach,” Pichai said in the Monday blog post.
Google has used AI tech to help improve search results for years. Its language algorithms parse peoples’ questions and queries and make guesses at what information would be most helpful. That’s why Google can easily tell you’re looking for “Sabrina the Teenage Witch” when you type in “TV show about a witch with a talking cat,” or know you’re looking for durians when you type in “big spiky fruit.” But chatbots like ChatGPT or Bard actually generate their own text based on all the information they’ve been trained on, so Google can create completely new pieces of content to help answer search queries.
The example the company gave in its blog post was a user asking Google search whether the piano or guitar are easier instruments to learn, and how much practice time each takes. The bot returned a three-paragraph answer, similar to what a music blog written by a real person may have provided in the past.
Google has been accused of stealing internet publishers’ content for years, such as using snippets of news articles in search results or pulling information from Wikipedia that it displays directly in search results rather than just providing links to the original content. But the use of large language models, which are trained on huge amounts of internet content, including copyrighted writing and news articles, is already intensifying this debate. A group of artists have sued Stability AI, an AI company that allows users to generate images, for copyright infringement because some of their images were allegedly used to train the software.
Still, companies big and small are charging ahead on the tech.
On Tuesday, Microsoft will hold an event that is widely expected to showcase how they will deploy technology from OpenAI in their own products. The company hasn’t confirmed the details of the event, but OpenAI CEO Sam Altman tweeted a photo of him and Nadella together on Monday, saying “excited for the event tomorrow.” Microsoft’s Bing search engine has long lagged far behind Google’s. Both companies have sold AI tools through their cloud software businesses, an area where Microsoft leads Google.
The technology powering Google’s Bard chatbot is not brand-new. The company showed off the chatbot tech, known as LaMDA, in 2021 at its annual developer conference. It stressed that the bot could be used for educational and scientific purposes, like helping kids learn about the solar system. Last year, the company fired one of its engineers after he spoke out about his beliefs that LaMDA had become sentient.
Throughout that time, Google has kept the technology internal and under wraps, but the hype and energy around generative AI has now pushed the company to move faster and publicize it.