Digital Safety
Independent AI safety assessment research
The Digital Safety Research Institute prevents harm and saves lives through independent AI safety research and engineering, developing methodologies, tools, and infrastructure to assess AI systems as technology evolves. Through research, open-source tools, partnerships, and education, the institute advances a scalable ecosystem for monitoring and mitigating risks associated with artificial intelligence.
Discover how our researchers are working to reduce AI risks and discover more Stories of Impact from all of our research institutes.
Agile AI assessments:
Research methods that help evaluators assess AI systems quickly and effectively.
Safety-focused research:
Developing tools to identify harmful, unsafe, and misleading AI outputs.
Open-source innovation:
Publishing assessment platforms and tools for broad community adoption.
Evaluation partnerships:
Collaborating with assessment organizations to scale AI safety testing.
Research challenges:
Competitive evaluations that accelerate breakthroughs in AI safety, media integrity, and risk assessment.
Building the Future of AI Safety Evaluation
The Digital Safety Research Institute combines scientific research, engineering, and collaboration to create practical AI safety assessment capabilities. Its work helps researchers, evaluators, policymakers, and organizations better understand and manage emerging AI risks.
Open scientific contributions
The Digital Safety Research Institute advances AI safety assessment research through scientific publications, research challenges, grants, and community engagement that shape the future of trustworthy AI evaluation.
Assessment tools for adoption
The Digital Safety Research Institute develops and shares open-source tools that organizations can use, assess, adapt, and integrate into their own AI safety evaluation programs.
Digital Safety Key Research Areas
AI chatbot safety assessment
The Digital Safety Research Institute develops verification and moderation methodologies that help determine whether AI chatbots respond correctly, safely, and reliably to user prompts. The research supports scalable testing of conversational AI systems and their outputs.
AI-generated media analysis
The Digital Safety Research Institute creates methodologies and tools to identify indications that media has been generated or edited by artificial intelligence. This work supports the detection of manipulated content and strengthens trust in digital information.
Agentic AI evaluation
The Digital Safety Research Institute is advancing future capabilities for creating, verifying, and validating AI agents. The goal is to improve assessment accuracy and accelerate the pace of AI safety evaluations as autonomous systems become more common.
AI risk intelligence & education
The Digital Safety Research Institute supports AI risk assessments, stakeholder education, and research infrastructure that enable organizations to better understand evolving AI threats and deploy safer technologies.
Digital Safety
News & Featured Updates
Stay up to date on the latest news and research from the Digital Safety Research Institute.
Stay informed
Sign Up for Updates
Knowledge and awareness about potential hazards associated with scientific innovations help in minimizing risks and promote best practices that improve safety. Sign up to learn more about our initiatives, partnerships, upcoming events, and free resources.
By submitting this form, I am opting in to receive communications from UL Research Institutes (ULRI) containing news and updates about its work and initiatives. I understand that I can manage my preferences at any time and agree to ULRI’s online policies.
Partnerships
Everyone benefits when world-class institutions work together to take on big challenges. Join us.
Grants
We’re funding scientific research to make the world safer, healthier, and more sustainable. Learn about funding opportunities.
Careers
Work to make the world a safer place. Start a rewarding career at UL Research Institutes with other socially minded colleagues focused on innovation, collaboration, and ethics.