Research Scientist, AI Safety & Harmful Manipulation

DeepMind Technologies Limited; Mountain View (CA); In loco; USD 174.000 - 252.000; Tempo pieno

Descrizione del lavoro

DeepMind Technologies Limited is seeking researchers within the Responsible Development and Innovation (ReDI) team to explore Harmful Manipulation risks and develop evaluations for AI-driven manipulation.

You will bridge research and practice, conduct technical investigations on model interactions with users, and publish novel methodologies while informing mitigation and governance decisions. This role emphasizes safety, ethics, and practical deployment considerations.

Competenze

• PhD or equivalent in a related field.
• Experience designing evaluations or studies on technology manipulation, persuasion, or related social phenomena.

Mansioni

• Design and implement evaluations identifying Harmful Manipulation-related capabilities and risks; publish on novel methodologies.
• Collaborate across teams to conduct holistic risk assessments and verify deployment safeguards for high-capability models.
• Assist mitigation efforts by providing novel and actionable insights across Google teams.
• Engage with external researchers and stakeholders, including governmental entities and academic communities.

Conoscenze

Evaluation design

Safety research

Formazione

PhD or equivalent in related field

Descrizione del lavoro

DeepMind Technologies Limited is seeking researchers within the Responsible Development and Innovation (ReDI) team to explore Harmful Manipulation risks and develop evaluations for AI-driven manipulation.

You will bridge research and practice, conduct technical investigations on model interactions with users, and publish novel methodologies while informing mitigation and governance decisions. This role emphasizes safety, ethics, and practical deployment considerations.