January 22nd 2024
AI Safety Institute (AISI) now recruiting
First ever state-backed organisation to focus on advanced AI safety seeks a range of skills and experience to join its expert team.
The Artificial Intelligence Safety Institute (AISI) was established late last year by the Department of Science, Innovation and Technology (DSIT) to examine, evaluate and test frontier AI models, and research how to keep society safe given the fast and unpredictable development of AI. This, of course, follows discussions and commitments made at the AI summit held in Bletchley at the start of November.
Photo by julien Tromeur
This work is being conducted in the public interest and findings will be shared openly to enable an effective global response to both the opportunities and risks presented by AI. The hope is to build a body of evidence that can lay the foundations for technically grounded international governance in this area.
AISI is now actively recruiting experts in a range of areas to help carry out this important work. These give us an insight into how the government is looking to tackle these areas.
Job adverts have been published for roles including:
Cyber Misuse Evaluations Lead who will build and lead a team employing techniques including threat modelling and cyber ranges to test cyber capabilities of frontier systems.
Loss of Control Evaluations Lead who will build and lead a team to evaluate AI capabilities that are precursors to extreme harms from loss of control, with a focus on autonomous replication and adaptation, and uncontrolled self-improvement.
Safeguard Analysis Lead to build and lead a team at the intersection of machine learning (ML) and security, to better understand how the safety and security components of frontier AI systems stand up to a range of threats.
Research Engineers who will develop tools and methods for testing AI systems, pushing the frontier of understanding and mitigations across all research teams at AISI.
Research Scientists to lead and contribute to projects integrated into AISI’s evaluation suite, evaluating frontier model capabilities and safeguards, as well as more speculative work aimed at mitigations and system understanding.
Chief Information Security Officer to build and lead a team to strengthen AISI’s cyber resilience and forge key partnerships with top AI firms and other government departments.
Software Engineer to create expert interfaces and accessible red-teaming frontends, providing fast and secure inference channels for internal and external models, as well as ML ops capabilities for hosting and fine-tuning AISI’s own models.
Head of Engineering to lead software engineers and research engineers.
Frontend Developer to blend web development, UX engineering and data visualisation to provide inference channels, facilitate hosting our own models and create expert interfaces for evaluations development.
UX Engineer to implement user interfaces that form part of AISI’s cutting-edge evaluations platform.
Applications can be made through the Civil Service Jobs website. AISI also welcomes expressions of interest from those keen to join the team more generally.
In related news:
Leave a Reply