Google appointed Josh Woodward to lead the Gemini app in April 2025 while he retained his role running Google Labs, putting its consumer AI assistant and experimental products unit under the same executive.
Google said in October that Gemini had reached 650 million monthly active users, up from about 350 million in March. The company also said AI Overviews, its AI-generated summaries in Search, reached about 2 billion monthly users.
Woodward replaced Sissie Hsiao, who had led the chatbot effort originally called Bard. In a staff memo announcing the change, Google DeepMind CEO Demis Hassabis said the move would “sharpen our focus on the next evolution of the Gemini app.”
Image tools and new models
Google introduced its Nano Banana image-generation and editing model in the Gemini app in late August 2025. The tool lets users combine photos and edit images. Google said the update was designed to preserve the likeness of people and pets across edits.
Gemini reached No. 1 on Apple’s U.S. App Store on September 12, displacing ChatGPT, according to app intelligence firm Appfigures. Google later said more than 5 billion images had been generated in the Gemini app. It also began extending Nano Banana to other products, including Google Lens, Circle to Search and NotebookLM.
At a November all-hands meeting, Google’s AI infrastructure chief Amin Vahdat described the demand on the company’s custom tensor processing units by saying, “Our TPUs almost melted.”
Google introduced Gemini 3 in November 2025, making Gemini 3 Pro available in preview across products including the Gemini app, AI Studio and Vertex AI.
The releases came alongside increased spending plans. In its October earnings report, Google’s parent company, Alphabet, forecast 2025 capital expenditures of $91 billion to $93 billion, up from its previous forecast of $85 billion.
Concerns about generated images
Google released Nano Banana Pro in November 2025. Built on Gemini 3 Pro, the image-generation and editing model subsequently drew criticism over its depictions of humanitarian aid in Africa.
Arsenii Alenichev, a researcher at the Institute of Tropical Medicine in Antwerp, said he observed “white saviour bias” and an association between dark skin and poverty in generated images. He also said the tool added charity logos he had not requested. Examples he shared showed women wearing charity-branded T-shirts surrounded by Black children.
Kate Hewitt, director of brand and creative at Save the Children UK, said, “These AI-generated images do not represent how we work.” She also expressed concern about third parties using the charity’s intellectual property for AI-generated content.
A Google spokesperson said in response: “At times, some prompts can challenge the tools’ guardrails and we remain committed to continually enhancing and refining the safeguards we have in place.”
Woodward’s work at Google Labs
Woodward, a 42-year-old Oklahoma native, began at Google as a product marketing intern in 2009 and joined Google Labs in 2022. His work there included overseeing the development of Project Tailwind, which became NotebookLM.
NotebookLM began rolling out in July 2023 to a small group of U.S. users. It uses selected documents to generate summaries and answer questions with citations to the underlying material. The product subsequently expanded to include audio and video features.
Woodward recruited author Steven Johnson as a part-time visiting scholar in 2022. Johnson joined full time in May 2023 and helped prepare a demonstration of the project for Google’s I/O developer conference.
In a March 2025 podcast conversation with Sequoia partners, Woodward said Labs aimed to move from an idea to users’ hands in 50 to 100 days. He described small teams developing products while working with Google DeepMind and other parts of the company.
Woodward has also used public feedback to address individual product issues. In October, he announced on X that Gemini users could change models during a conversation without starting over.
