Google consolidates teams with aim to create AI products faster

Alphabet chief executive Sundar Pichai announced changes to Google’s workplace teams structure, saying the moves will help the company develop artificial intelligence (AI) products and services faster and more efficiently.

In a note to employees, also published on Thursday in a blog post, Pichai said the models, research and responsible teams would be consolidated under the company’s flagship AI division, Google DeepMind.

To accelerate work on Google’s AI models – Gemini and Gemma – employees working on the technology across Google Research and Google DeepMind will unite as one team, which will also consolidate the expensive computing power needed to train and build the systems under one arm of the company. Responsible AI teams across the business will also move under Google DeepMind.

A new unified Platform and Devices team, meanwhile, will bring together efforts across hardware, software and AI teams at Google – including those working on products such as Android, Chrome, search and photos, Pichai said. The group will also include employees working on computational photography and on-device AI features like Google’s recently launched “circle to search” AI tool it announced in partnership with Samsung Electronics.

The changes, Pichai wrote, “will help us work with greater focus and clarity toward our mission”.

While Google has intensified its work on generative AI to catch up to the allied effort of Microsoft and ChatGPT creator OpenAI, which many perceive as being more advanced, the company also has been shifting priorities and lowering costs. For the past few months, that has resulted in a series of cascading job reductions, creating a grim new reality of insecurity for employees. In January, the company cut hundreds of people working on its digital assistant, hardware and engineering teams.

At the same time, Google framed the reorganisation as necessary to more sharply focus on launching AI tools and services. Earlier this year, the company rolled out a new version of its powerful AI model, Gemini 1.5 Pro, which it said could handle a larger amount text, video and even audio outputs compared with the competition. It also rebranded its chatbot as Gemini and released a more open large language model to help the company gain favour with the open-source community.