One of Google’s efforts to maintain pace with the technological titans is consolidating its AI research divisions. The company recently announced that it was combining its Advanced Technologies and Projects (ATP) division with Deep Learning, Computer Vision and Natural Language Processing research groups. This move will help the company better organize its resources and make it more efficient in developing groundbreaking AI technologies. In addition, the creation of a new research lab focused on Advances in Personal
Google DeepMind is a new unit made up of the DeepMind team and the Google Brain team from Google Research. This new unit will work together to deliver AI research and products to customers across Google’s product areas.
Google DeepMind has been in the news a lot recently due to their impressive progress in AI development. The company has been creating some of the most advanced artificial intelligence (AI) technology in the world, and they have now announced that they are creating a new scientific board to oversee research progress and direction of the unit. This board will be lead by Koray Kavukcuoglu, VP of research at DeepMind, and Eli Collins, VP of product at Google Research will join Google DeepMind as VP of product while Zoubin Ghahramani will become a member of the research leadership team, reporting to Kavukcuoglu.
Google’s announcement that it has formed a new unit called DeepMind ― led by artificial intelligence pioneer Demis Hassabis ― signals the company’s intent to remain at the forefront of AI technology. DeepMind will be tasked with developing “the next wave of world-changing breakthroughs,” and Hassabis is adamant about using responsible methods to achieve this goal.
In April of 2018, Google announced they had acquired an AI startup called DeepMind for $500 million. DeepMind, which is co-founded by Shane Legg and Mustafa Suleyman, developed a number of AI programs that have made headlines in the past few years. One of their most significant achievements was developing AI that could beat can human professional Go players. Additionally, their research into protein structure prediction led to the creation of a computer program that can identify Structure assignments with high accuracy. With Google’s resources and ongoing commitment to advancing AI research, it is likely that further feats will be achieved by DeepMind in the years to come
Amidst Google’s impressive array of research projects lies a group of computer scientists who are devoted to the development and exploration of artificial intelligence. Known as Google Brain, this small team consists of experts in machine learning, data mining, and computer vision. Their work has been instrumental in the development of artificial intelligence technologies such as TensorFlow and Google Translate.
DeepMind has long been a strictly independent company, but its close relationship with Google Brain has occasionally sparked controversy. For example, in December of last year, DeepMind was accused of sharing confidential data with Google Brain without proper consent from the data owners. And just this month, reports surfaced that the two companies had clashed over how best to use AI technology for military purposes. While these are isolated incidents and DeepMind and Alphabet are mostly on good terms now, they underscore the importance of autonomous research organizations maintaining their independence even as they work closely with corporate partners.
Senior managers at DeepMind reportedly tried to negotiate for years with Google for more autonomy, seeking an independent legal structure for the sensitive research they do. The talks were called off in May 2021. In a letter obtained by The Information, employees describe how senior managers “intensified their efforts” over the summer of 2020 before talks broke down in May. According to the letter, Google’s leadership team stated that it would only consider allowing DeepMind to operate as an apatride company if it abandoned its work on artificial intelligence and Cognitive Science altogether.
DeepMind has made a number of acquisitions in recent years, including artificial intelligence companies Merlin Technologies and Nervana Systems. These acquisitions have helped DeepMind to continue its growth in the AI field. However, DeepMind has also struggled to break even with its growing expenses, which include the compute necessary to train massive AI models and a large headcount (around 1,000 employees). Alphabet was forced to write off nearly £1.1 billion ($1.5 billion) of DeepMind’s debt in 2019. However, Deepmind is now profitable, reporting a modest £44 million ($60 million) profit in 2020, up from a £477 million ($650 million) loss in 2019. The majority of DeepMind’s revenue is dependent on licensing its technologies to other Alphabet groups like Waymo
DeepMind’s prowess in AI has made it a valuable addition to Alphabet, which is looking to overtake rivals like Microsoft and OpenAI in the race for revenue-generating AI. While DeepMind’s successes remain largely theoretical at this point, they could provide important guidance as Alphabet perfects its own artificial intelligence methods.
It seems that Google is working on a software application that can compete with OpenAI’s Gemini project. This comes as a surprise, as the recent AI product pushes from Alphabet have been met with poor reviews. It’s likely that this new development will help improve customer confidence in Alphabet’s AI ventures.
Google’s close collaboration with OpenAI on Bing Chat is alarming Microsoft, which is reportedly investing outsize resources into its own Magi search features with AI capabilities. As Google invests heavily in artificial intelligence, it may soon have a technological advantage over its competitors.