AI systems are increasingly deployed across diverse industries, including healthcare, education, agriculture, and peace and security sectors like policing and military applications. To ensure responsible deployment, AI designers, developers, and deployers must strike a balance between harnessing AI benefits and mitigating potential harm. This requires adhering to the principles of proportionality and 'do no harm', which advocate for implementing measures to limit or restrict the design, development, and deployment of technologies to what is necessary for achieving their intended goals while minimizing associated risks.

It is crucial for AI actors to center their practices on the foundational principle of 'do no harm', not only in high-risk technology contexts but also in high-risk environments. They should proactively mitigate risks and address harms caused by AI to prevent human rights abuses over time.