Research Scientist, Harmful Manipulation

On-site in London·Added yesterday (last checked 3 Nov)

Google

265 open roles

Apply

Our summary

Work mode: On-site, London

Requirements

What we're looking for

PhD degree in Psychology, Political Science, Computational Social Science, Computer Science, a related field, or equivalent practical experience.
Experience designing evaluations or studies to investigate the intersections of technology and manipulation, persuasion, social engineering, or related social phenomena.

Nice to have

Track record of driving complex projects, including in research settings.
Track record of publications in high impact domain-relevant conferences or journals (NeurIPS, CHI, FAccT, etc).
Ability to leverage AI tools to build and accelerate research pipelines or safety benchmark development.
Commitment to translating research insights into practical applications and working with cross-functional teams.

Pay & benefits

What you'll get

Individual pay is determined by factors including job-related skills, experience, and relevant education or training.

US: $174000 - $252000 (USD) + 15% bonus target + equity + benefits

The role

At DeepMind, the frontier of AI brings extraordinary opportunities and unprecedented responsibilities.

As part of the Responsible Development and Innovation (ReDI) team, you will be responsible for researching emerging risks within the Harmful Manipulation risk domain, developing and implementing novel evaluations to drive understanding of AI-driven harmful manipulation. Your work will range from conducting technical investigations, probing how DeepMind's most advanced AI models interact with and influence users, to informing mitigation and governance decisions, to publishing on novel methodologies and engaging with external stakeholders. You will have an opportunity to bridge research and practice, contributing novel approaches to help inform model development and deployment.

Artificial intelligence will be one of humanity's most transformative inventions. At Google DeepMind, we are a pioneering AI lab with exceptional interdisciplinary teams focused on advancing AI development to solve complex global challenges and accelerate high-quality product innovation for billions of users. We use our technologies for widespread public benefit and scientific discovery, ensuring safety and ethics are always our highest priority.

We are pushing the boundaries across multiple domains. Our global teams offer diverse learning opportunities and varied career pathways for those driven to achieve exceptional results through collective effort.

What you'll do

Partner cross-functionally to design and implement evaluations identifying Harmful Manipulation-related capabilities and risks; publish on novel methodologies and capabilities insights
Collaborate across teams to carry out holistic risk assessments, including determining if deployment safeguards are adequate for high-capability models.
Work across evaluation and mitigation teams to assist mitigation efforts, while collaborating across Google to provide novel and actionable insights.
Engage with external research community and stakeholders, including governmental entities, national safety institutes, third-party organisations (including the Frontier Model Forum), external subject matter experts, and academic communities.

Learn more about benefits

Learn more about benefits at Google. Note: By applying to this position you will have an opportunity to share your preferred working location from the following: New York, NY, USA; London, UK; Mountain View, CA, USA.

Hiring process

4 steps3 interviewsa task to do6–8 weeks in all

Apply online
Assessments
Online test · Not every role
You may be asked to take the Google Hiring Assessment (workstyle skills) and, for some roles, role-related assessments such as a coding exercise.
Recruiter conversations
With a recruiter · Phone call · 1–2 sessions
One or two shorter conversations assessing key skills, sharing information about the organisation and team, and asking about your preferences such as location.
Project work
In your own time · Not every role
For some roles, a small project before interviews: a case study, writing or code samples.
Interviews
With a rotating cast of Googlers
A panel of structured interviews, over video or in person, with a rotating cast of Googlers; open-ended role-related questions scored against the same rubric for every candidate, no brainteasers.
Decision and offer · With a recruiter

Structured interviews: every candidate for a role is scored against the same rubric; no brainteasers.
AI tools are not permitted during interviews.
For some roles, pre-selected interviews happen virtually before final interviews onsite.