Google is rushing to release its own artificial intelligence products in the wake of OpenAI’s ChatGPT. The search engine pioneer is working hard and fast on a “code red” effort to respond to ChatGPT with a large language chatbot and testing new ways to incorporate that AI-powered bot into search, according to a report from CNBC.
The new report backs up earlier news from the New York Times and elsewhere, which outlined a rapid re-alignment in Google’s priorities in direct response to the rise of ChatGPT. CEO Sundar Pichai reportedly re-assigned employees and “upended” meetings to boost the amount of resources going towards the company’s AI development.
CNBC’s Tuesday account offers further details. Google’s new chatbot, reportedly named “Apprentice Bard,” is based on the company’s pre-existing LaMDA (Language Model for Dialogue Applications) technology. The application looks and functions similarly to ChatGPT: Users input a question in natural language and receive a generated text response as an answer. But Apprentice Bard seemingly has a couple of important skills beyond what ChatGPT can do.
For one, it can draw on recent events and information, according to CNBC, unlike ChatGPT which is limited to online information from before 2021. And it may be better at achieving that elusive AI accuracy. For instance, LaMDA correctly responded to a math riddle that ChatGPT failed to grasp, as recorded in company documents viewed by CNBC.
Google leadership acknowledged to employees that the company has much more riding on the results accuracy of its future AI endeavors than other competitors in the sector. Google is, after all, the internet’s preeminent source for information. People come to the search engine to find facts—even if it might not always lead them there. The tech giant carries more “reputational risk,” if it were to provide users incorrect info, and therefore is progressing “more conservatively than a small startup,” Google’s AI chief, Jeff Dean, told employees in an all-staff meeting, according to CNBC. But moving cautiously doesn’t mean the company isn’t moving at all.
Google is reportedly already looking for ways to bring AI into search. The company is internally testing including Apprentice Bard on a version of their main page. In one trial version, a user inputs a query in the search bar, as they do now, but the top result is a human-like, written response displayed in a gray bubble directly below the bar. Then, beneath that, the results page suggests a set of AI-generated follow-up questions related to the first. And finally, below the robot chat, the rest of the search results are listed, according to CNBC’s description.
Gizmodo reached out to Google to confirm the details of the CNBC report and didn’t receive a direct response to our questions. Instead, company spokesperson Lily Lin sent the following statement:
We have long been focused on developing and deploying AI to improve people’s lives. We believe that AI is foundational and transformative technology that is incredibly useful for individuals, businesses and communities, and as our AI Principles outline, we need to consider the broader societal impacts these innovations can have. We continue to test our AI technology internally to make sure it’s helpful and safe, and we look forward to sharing more experiences externally soon.
Important to note though: Even if CNBC’s account is a completely accurate reflection of Alphabet-owned Google’s internal tests, that doesn’t mean any final AI-product(s) that the company releases will exactly match this early report. And, even though some analysts and pundits have forecast that the advent of ChatGPT and similar AI is a death knell for traditional internet search, Google’s standard search engine probably isn’t going anywhere for now.
Getting an AI to answer one math riddle correctly doesn’t exactly prove that Google has cracked the technology’s persistent problem of lying and fabricating incorrect answers. Users probably aren’t looking to spend their time fact-checking search results.
However, with Microsoft’s recently renewed multi-billion dollar partnership with OpenAI announced, it makes sense that Google would be on high alert. Microsoft owns Bing, one of Google’s biggest competitors on the search engine market. Thanks to the new deal, Microsoft is reportedly set to add an updated version of ChatGPT into Bing.