Google was reportedly planning a broad wave of artificial intelligence products for 2023, with a chatbot-enabled search engine among its most prominent projects. The reported effort followed ChatGPT’s rapid success and came as Google considered how to bring more conversational features into a service central to its business.
ChatGPT raises the stakes for search
According to The New York Times, Google founders Larry Page and Sergey Brin joined several meetings with company management in December 2022 to discuss its AI strategy. The report said the discussions covered chatbot search and a plan to introduce 20 AI products during 2023. The founders, who stepped away from day-to-day operations in 2019, were reportedly invited by CEO Sundar Pichai.
ChatGPT’s ability to respond to questions in conversational form suggested a possible alternative to some search queries, although the source notes that many questions remained unanswered. That possibility made the technology a challenge for Google’s search business and a reason to consider how chatbot features might fit into its own search engine.
Competition with Microsoft added another dimension. The two companies compete in cloud computing, a market the source describes as lucrative and driven primarily by price. Exclusive AI features could give a company an advantage, making product plans relevant beyond search and consumer software.
A wide range of products was reportedly in the works
The New York Times report, citing slides and two anonymous sources familiar with the project, said Google expected to present products at its I/O conference in May. One reported project, “Image Generation Studio,” would support AI image generation and editing. Google’s Imagen model had been unveiled the previous year and was in beta testing in the U.S.; Re-Imagen was described as a way to customize image models.
Other reported plans touched several Google products and services:
- A new version of the A.I. Test Kitchen environment.
- A YouTube green screen feature for custom backgrounds called “Shopping Try-on.”
- A wallpaper generator for Pixel smartphones.
- An app called “Maya” for viewing shoes in 3D.
- Software that creates summaries of videos.
The plans also included tools for developers. MakerSuite was meant to help them build AI prototypes for the internet browser, while Colab and Android Studio were reportedly set to generate, complete and repair code. A planned coding model called PaLM-Coder 2 was designed to generate and complete code and appeared to be based on Google’s PaLM language model.
Speed had to be balanced with reliability
The reported product push was accompanied by changes to how Google handled AI development. Pichai was said to have sped up product approval and created a faster path for AI projects called “Green Lane.” The report also described working groups intended to release products sooner and review them more independently.
Google was reportedly considering a broader recalibration of its willingness to take risks when launching AI technology. But the proposed chatbot search product was expected to face particular scrutiny. The company was said to prioritize accurate answers and safety, in part because ChatGPT had drawn criticism over the reliability of its responses.
Controls were also part of the reported plan. Google was supposed to block certain words to reduce hate speech, while the source said the bar would be lowered for some other upcoming products. That points to different release considerations across the portfolio: chatbot search, which answers questions directly, was described as receiving extra caution.
Public concerns could shape the rollout
The report linked the caution around chatbot search to political concerns as well as product quality. Google’s leadership reportedly anticipated greater pressure for AI regulation amid worries about misinformation, harmful content, bias and copyright. Those concerns help explain why moving quickly did not mean treating every product launch the same way.
Google spokeswoman Lily Lin told the NYT: “We continue to test our A.I. technology internally to make sure it’s helpful and safe, and we look forward to sharing more experiences externally soon,” The quote reflects the company’s stated approach at the time: continue internal testing while preparing to show more AI products outside the company.
Taken together, the reported plans described a company responding to a fast-moving competitor while expanding AI across search, images, video, mobile devices, cloud services and coding tools. The chatbot search proposal stood apart because it combined the prospect of a new way to find information with concerns about whether its answers would be correct and safe.