Google AI Responsibility Shift: Powerful DeepMind Shake-Up Moves 90-Person Team to Global Affairs
Google AI Responsibility Shift Brings a New Chapter for DeepMind
The Google AI responsibility shift has become the latest major change inside Google’s artificial intelligence business. The company is reportedly moving a roughly 90-person AI responsibility team out of Google DeepMind and into Google’s Global Affairs organization, which handles areas such as public policy and government relations.
According to a Wall Street Journal report cited by several news organizations, the move is expected to be completed at the beginning of September. The team works on some of the most important questions surrounding advanced artificial intelligence, including model safety, possible risks and the wider effects of AI on society.
The Google AI responsibility shift comes only weeks after a wider leadership change at Google DeepMind. That earlier restructuring placed more focus on the company’s Gemini AI models and on moving artificial intelligence products forward at a faster pace.
For Google, the latest change appears to be part of a broader effort to create a more centralized structure around AI safety, policy and responsibility.
At the same time, the move has attracted attention because the responsibility team has traditionally worked closely with researchers building Google’s AI models.
What Is the Google AI Responsibility Team?
The Google AI responsibility shift is easier to understand when looking at what the team actually does.
Google DeepMind’s Responsibility team works on the safe and responsible development of artificial intelligence. According to Google DeepMind, the team includes experts in areas such as AI governance, assurance evaluations and risks involving chemical, biological, radiological and nuclear threats.
The team also looks at the possible effects of AI systems on people and society.
This work can include testing AI models, studying potential misuse and helping research and product teams understand risks before new systems are released.
Google DeepMind says its responsibility teams work with internal groups and the wider AI community to embed ethics into AI development and test systems carefully.
Therefore, the Google AI responsibility shift is not simply an office move. It changes where an important group of AI experts sits inside the company’s larger structure.
Why Google Is Making the Change
The exact reasons behind the Google AI responsibility shift have not been fully explained publicly. However, an internal email cited by the Wall Street Journal reportedly described the change as part of moving Google DeepMind toward a more traditional structure in which some central teams sit within Google itself.
Helen King, a Google DeepMind vice president, reportedly told employees that the organization was moving toward a more typical structure with central teams operating within central Google.
Google has also said that the change is intended to strengthen its AI responsibility efforts.
This is important because AI safety is no longer only a research issue. As AI becomes part of search, software, business tools and consumer products, questions about policy and public impact have become much more important.
The Google AI responsibility shift could therefore be aimed at bringing safety and policy work closer to teams that deal directly with governments, regulators and other outside groups.
Why Some Employees Are Concerned
The Google AI responsibility shift has reportedly raised concerns among some DeepMind employees.
The main concern is about independence.
When a responsibility team sits inside a research organization, it may have a close connection with the scientists who build AI systems. This can make it easier for safety experts to raise concerns during the development process.
Moving the team into Global Affairs could create a different relationship.
Some employees reportedly worry that the team may become less independent or may have less direct involvement in examining Google’s newest AI models.
However, this does not automatically mean that AI safety work will become weaker.
Google could also be trying to give the team a wider role across the company.
The final impact will depend on how the new structure works in practice.
Google Has Changed Its AI Structure Before
The Google AI responsibility shift is not the first time Google has changed the way its AI safety teams operate.
In 2024, Google announced changes to its Responsible AI teams. The company said it was moving Responsible AI teams in Research into Google DeepMind so they could be closer to the teams building and scaling AI models.
Google said at the time that the move was designed to create clearer responsibility and accountability and strengthen the feedback loop between AI models, products and users.
That earlier decision makes the current change especially interesting.
The company previously emphasized the value of placing responsibility experts close to AI researchers. Now, part of that work is reportedly being moved into a central Google organization.
This shows how quickly AI organizations are changing as companies learn what structure works best for them.
The Timing Is Important
The Google AI responsibility shift comes during a major period of change for Google DeepMind.
Earlier in August, Alphabet announced a leadership overhaul at Google DeepMind. Demis Hassabis moved from the CEO role into a chairman position, while Koray Kavukcuoglu took over as head of the AI division.
The restructuring also came as Google continued working to strengthen its Gemini AI models and compete with companies such as OpenAI and Anthropic.
Reuters reported that Google’s AI organization had been facing pressure to move faster in the competitive AI market. The company has been working to improve Gemini while also dealing with internal organizational challenges.
Therefore, the Google AI responsibility shift is happening as the company is trying to balance two important goals: moving quickly in AI and maintaining responsible development.
Google Wants to Move Faster in AI
The Google AI responsibility shift is also connected to Google’s larger AI strategy.
The artificial intelligence market has become extremely competitive. Companies are investing billions of dollars in computing power, research and AI products.
Google has powerful AI systems, large data resources and its own computing infrastructure. However, the company is also facing strong competition from other technology firms.
Reuters reported earlier this month that Google leaders have been pushing the company to move faster in the AI race. The report described a broader reorganization aimed at improving Gemini’s development and commercial progress.
This creates a difficult balance.
AI companies want to release better systems quickly. At the same time, advanced AI models need careful testing before they are widely used.
The Google AI responsibility shift could be one way Google is trying to organize these responsibilities more clearly.
AI Safety Is Becoming More Important
The Google AI responsibility shift also highlights how much AI safety has changed in recent years.
When AI systems were less capable, many safety questions were mainly discussed inside research departments.
Today, AI models can generate text, images, software code and other content. They are also being used in workplaces, education, customer service and many other areas.
As a result, companies need to think about accuracy, privacy, misuse, security and social impact.
Google DeepMind says its approach includes anticipating and evaluating a broad range of AI risks. Its Responsibility and Safety Council reviews research, projects and collaborations against Google’s AI principles.
The company also has an AGI Safety Council focused on possible extreme risks from future advanced AI systems.
These groups show that AI responsibility is becoming a much wider field.
What Does Global Affairs Do?
To understand the Google AI responsibility shift, it is also useful to understand Google’s Global Affairs organization.
Global Affairs deals with issues involving governments, regulators, public policy and other external stakeholders.
AI companies increasingly need teams that can explain their technology to policymakers and respond to new laws.
Governments around the world are creating rules for artificial intelligence. These rules can cover areas such as privacy, safety, transparency and the use of AI in important services.
By placing a responsibility team closer to Global Affairs, Google may be looking to connect technical AI safety work with policy discussions.
This could help the company communicate more effectively with regulators and other groups.
Could the Move Improve Coordination?
The Google AI responsibility shift could have some potential benefits.
First, it could make it easier for safety experts to work across different parts of Google instead of focusing mainly on DeepMind.
Google has AI systems across Search, Android, Cloud, Workspace and other products.
A central responsibility team could potentially look at AI risks across all these products.
Second, the team could have a stronger connection with policymakers.
As AI rules develop in the United States, Europe and other regions, companies need people who understand both technology and policy.
Therefore, the move could help Google coordinate its AI responsibility work more broadly.
Could There Be Downsides?
At the same time, the Google AI responsibility shift could create challenges.
The biggest question is how closely the team will continue to work with AI researchers.
AI safety experts need access to models and development teams so they can identify potential problems early.
If the new structure creates too much distance between responsibility experts and researchers, some important feedback could become slower.
That is why the practical details of the transition matter more than the organizational name itself.
If the team remains deeply involved in AI testing and development, the change may have limited impact on day-to-day safety work.
If its role becomes mainly focused on policy, the change could be more significant.
Google Says Responsible AI Remains Important
The Google AI responsibility shift should not be seen as evidence that Google is abandoning AI safety.
Google DeepMind continues to publicly describe responsible AI development as a central part of its work.
Its official Responsibility and Safety page says the company wants to build AI responsibly and evaluates systems against a broad range of risks.
Google DeepMind’s careers page also describes its Responsibility team as a group that works on AI ethics, system testing and societal impacts.
Therefore, the key issue is not whether Google will continue talking about responsible AI.
Instead, the important question is how responsibility work will be organized and how much influence it will have during AI development.
A Bigger Change Inside DeepMind
The Google AI responsibility shift also reflects the gradual transformation of Google DeepMind.
DeepMind was originally known as a research-focused AI lab. Google acquired the company in 2014 and later merged it with Google Brain.
Since then, the organization has become much more closely connected to Google’s commercial products.
The latest restructuring suggests that this process is continuing.
Reuters reported that several nontechnical teams were also being moved from DeepMind into Google’s broader corporate structure.
That means the responsibility team is not necessarily being moved in isolation.
Instead, it appears to be part of a wider effort to organize DeepMind around Google’s larger AI strategy.
What This Means for Gemini
The Google AI responsibility shift could also affect how Google’s Gemini models are developed and reviewed.
Gemini is now one of Google’s most important technology projects.
The company wants Gemini to compete with leading AI models from other companies while also using it across many Google products.
As Gemini becomes more capable, safety testing becomes increasingly important.
The responsibility team could play a role in checking how models behave, identifying risks and helping Google decide how systems should be released.
The new organizational structure will determine how that work connects with Gemini’s development.
What Employees and Users Will Watch
The Google AI responsibility shift will likely be watched closely by both employees and the wider AI community.
Employees will want to know whether their daily work changes and whether they continue to have access to researchers and model-development teams.
AI researchers may also watch whether the new structure changes the way safety concerns are raised.
Users, meanwhile, are likely to care about the practical outcome.
If Google’s AI products become more useful while remaining safe and reliable, the organizational change may have little importance to everyday users.
However, if the move changes how AI risks are evaluated, it could become more important over time.
What Happens Next?
The Google AI responsibility shift is expected to take effect around the beginning of September, according to reports citing Google’s internal communication.
The next few months should provide a clearer picture of how the team operates under Global Affairs.
Google will likely continue developing Gemini and other AI systems at a rapid pace.
At the same time, regulators and governments will continue creating new rules for AI.
This means the need for cooperation between technical experts, safety researchers and policy specialists is likely to increase rather than decrease.
The new structure could therefore become an important test of how large technology companies organize AI responsibility as the technology grows.
Conclusion
The Google AI responsibility shift is another significant change in Google’s fast-moving AI organization. The company is reportedly moving a roughly 90-person responsibility team from Google DeepMind to its Global Affairs organization, with the transition expected to be completed in early September.
The team works on important areas such as AI safety, model testing, governance and the wider effects of artificial intelligence on society.
Some DeepMind employees have reportedly expressed concerns about the team’s independence. However, Google says its goal is to strengthen and organize AI responsibility work more effectively.
The change also comes during a much larger restructuring at Google DeepMind as the company works to make Gemini more competitive and move AI development faster.
Ultimately, the success of the Google AI responsibility shift will depend on how the new structure works in everyday practice.
A strong AI organization needs both innovation and careful review. These goals do not have to conflict with each other. With clear communication, good testing and cooperation between researchers and policy experts, companies can continue developing useful AI while also taking reasonable care of the people who use it.
For now, the change is best viewed as another step in Google’s ongoing effort to find the right balance between fast AI development, responsible technology and a changing policy environment.
Sources: Reuters, The Wall Street Journal reporting as cited by The Times of India and QZ, Google DeepMind’s official Responsibility and Safety information, and Google corporate publications. The article is based on publicly reported information available on August 27, 2026.
