A2IS2E
Sectors

Third party evaluation and audit

We help organisations that serve the public prove their commitment to AI safety. Our independent evaluation and audit confirms that your AI systems meet shared standards of dignity, inclusion, and harm prevention, grounded in ubuntu.

Talk to usSee our services

01

Public Sector

We test citizen-facing AI used by government agencies, regulators, and public service providers. Our services include:

  • Red teaming chatbots that answer citizen questions about benefits, immigration, or taxes
  • AI procurement checklist development
  • Training public sector teams to conduct internal red teaming
  • Fractional red team support for ongoing public service AI projects
  • Pro bono evaluation when a public concern has been raised

02

Public Health

Public-facing health AI must be safe, accurate, and accessible. Our services include:

  • Testing symptom checkers and triage chatbots for dangerous advice
  • Red teaming clinical decision support tools used by frontline workers
  • Testing AI health data analysis and reporting tools for errors or privacy leaks
  • Training public health teams and community health workers in red teaming

03

Biosecurity

AI in biosecurity can be of great benefit to society, but it can also introduce new risks. Our services include:

  • Evaluating AI systems used in public health communication for misinformation risks
  • Probing tools that support lab safety, pathogen research, or biological threat screening
  • Assessing dual-use risks in AI models accessible to the public
  • Training biosecurity and public health teams to identify harmful model behaviors

04

Civil Society

We continue to serve non-profits, human rights organisations, media outlets, and community groups. Our services include:

  • Pro bono red teaming where public concern has been raised and budget is limited
  • Training civil society teams to evaluate AI, Algorithmic, and Autonomous systems for harms

Pro Bono Across All Sectors

Where a public concern has been raised about an AI system in any of these sectors, and the responsible organisation lacks budget, A2IS2E can provide pro bono red teaming and evaluation. We prioritize cases involving vulnerable populations, essential services, and high public impact.

Report AI HarmRequest pro bono support