TL;DR:
- Google has combined its AI research divisions into Google DeepMind, a unit that will work closely with various Google product areas to drive AI research and product development.
- A new scientific board will be established to oversee the progress and direction of the unit, led by Koray Kavukcuoglu, VP of Research at DeepMind.
- Jeff Dean will serve as the Chief Scientist for both Google Research and Google DeepMind, while Google Research will remain an independent division focusing on advancing computer science.
- DeepMind was founded in 2010 and acquired by Google for $500 million in 2014. It has made headlines for its AI advancements, including the development of AI that can beat professional Go players and predict complex protein structures.
- Tensions have been reported between DeepMind and its parent company, Alphabet, with senior managers attempting to negotiate for more autonomy in the past.
- DeepMind has faced challenges with its growing expenses and is heavily reliant on licensing its technologies to other Alphabet groups.
- The software engineers from the Google Brain group are collaborating with DeepMind employees to create a software that can compete with OpenAI, called Gemini.
- Google is also investing heavily in Magi, a collection of AI-powered search features, in response to Microsoft’s close collaboration with OpenAI.
Main AI News:
The Google DeepMind unit, a collaboration between the DeepMind team and Google Brain team, was announced today by Google. In a blog post, the CEO and co-founder of DeepMind, Demis Hassabis, stated that the new unit would work closely with various Google product areas to drive AI research and product development. The formation of Google DeepMind will involve the creation of a new scientific board, responsible for overseeing the progress and direction of the unit.
Koray Kavukcuoglu, VP of Research at DeepMind, will lead the board, while Eli Collins, VP of Product at Google Research, will join as the VP of Product. Zoubin Ghahramani, lead of Google Brain, will also be part of the research leadership team, reporting to Kavukcuoglu. Additionally, Jeff Dean, one of the co-founders of Google Brain, will serve as the Chief Scientist for both Google Research and Google DeepMind, reporting directly to Google CEO Sundar Pichai.
Google Research will continue to operate as an independent division and will focus on advancing computer science in areas such as algorithms, privacy and security, quantum computing, health, climate and sustainability, and responsible AI. The division will report to James Manyika, SVP of Technology and Society at Google.
“With Google DeepMind, we aim to bring together our AI expertise and computing power, infrastructure, and resources to drive the next generation of AI breakthroughs and products across Google and Alphabet in a bold and responsible manner,” Hassabis wrote. “The work we will be doing as part of this new unit will create the next wave of world-changing breakthroughs.”
Hassabis co-founded DeepMind in 2010, which was later acquired by Google for $500 million in 2014. The lab has since made headlines for its AI advancements, including the development of AI that can beat professional Go players and predict complex protein structures.
The Google Brain team was formed a year after DeepMind in 2011 as a result of a part-time research collaboration between Jeff Dean, Greg Corrado, a Google researcher, and Andrew Ng, a Stanford University professor. The team achieved great success and created tools such as TensorFlow, an AI model creation framework, and improvements to Google Translate.
Despite the acquisition, DeepMind and Google Brain have mostly worked together seamlessly, though there have been reported tensions between DeepMind and its parent company, Alphabet. Senior managers at DeepMind attempted to negotiate for more autonomy for years and sought an independent legal structure for the sensitive research they conduct, but these talks were discontinued in May 2021.
DeepMind has faced challenges with its growing expenses, including the computational resources needed to train large AI models and a significant workforce of around 1,000 employees. The alphabet was forced to write off nearly £1.1 billion ($1.5 billion) of DeepMind’s debt in 2019. Although the company is now profitable, reporting a modest £44 million ($60 million) profit in 2020, up from a £477 million ($650 million) loss in 2019, its revenue is heavily reliant on licensing its technologies to other Alphabet groups, such as Waymo. Recently, DeepMind has gained increased importance within Alphabet as the company strives to compete with rivals like Microsoft and OpenAI in the race for revenue-generating AI.
Google Brain software engineers and DeepMind employees are reportedly working together to create a new software, called Gemini, to compete with OpenAI. This collaboration comes after recent AI product releases from Google, such as the chatbot Bard, received negative feedback and caused a decrease in Alphabet’s stock value. In response to Microsoft’s close collaboration with OpenAI on Bing Chat, an AI-powered chatbot integrated with the Bing search engine, Google is also reportedly investing heavily in Magi, a collection of AI-powered search features. Magi has a dedicated task force of over 160 people, established this year.
“We’ve been an AI-first company since 2016, as we see AI as the most impactful way to fulfill our mission,” wrote Pichai in a blog post published today. “The pace of progress in AI is faster than ever before. To ensure the responsible and bold development of general AI, we’ve created Google DeepMind to help us build more advanced systems safely and responsibly.”
Conlcusion:
The creation of Google DeepMind, a combination of DeepMind and Google Brain, is a strategic move by Google to bring together its AI expertise and resources to drive the next generation of AI breakthroughs and products. The formation of a new scientific board, led by Koray Kavukcuoglu, will oversee the progress and direction of the unit, while Jeff Dean will serve as the Chief Scientist for both Google Research and Google DeepMind.
The recent creation of Gemini, a software designed to compete with OpenAI, and Magi, a collection of AI-powered search features, highlights Google’s investment in AI technology and its efforts to compete with rivals like Microsoft and OpenAI. The AI market is expected to grow rapidly in the coming years, and Google’s latest move positions the company well to capture a significant share of the market.
