Summary of Guidance on Safe AI by the World Health Organization

Writer Profile Image
Ananya Avasthi
October 29, 2021
twitter iconfacebook iconlinkedin icon
copy url icon

The increasing use of AI-powered solutions has added value to traditional public health interventions. There is absolutely no doubt that AI has taken the world by storm and created a new way of working in different sectors.                                                                                                                                                                                                                      

WHO (World Health Organization) has taken a lead in researching potential health implementation with AI in certain regions and countries. A recent example, WHO is dedicated to establishing a foundation to boost the local impression of digital learning during public health emergencies. WHO aims to streamline the transcription and translation process by using machine learning algorithms. It is the wish of  WHO to assist the outcome of vulnerable populations and marginalized communities through using AI. WHO is currently developing guidelines for using AI to detect cervical cancer.  They have launched 2 projects on Early detection of breast cancer and AI and X-rays and COVID-19.

In certain countries, AI is being used to improve the speed and accuracy of diagnosis and screening for diseases. These steps include:

-Assisting with clinical care

-Boosting health research and drug development 

-To endorse diverse public health interventions, like disease surveillance, outbreak response, and health systems management.

AI can be used to educate and empower patients to take control of health needs and implement good habits for prevention as well. AI could have a deep impact on rural communities and countries with poor resources. The patients in these areas often have almost zero access to healthcare workers or medical professionals. To improve this situation WHO is taking a stand and implementing certain guidelines to benefit all people.


WHO has come up with the following principles as the foundation for AI regulation and governance to limit the risks and maximize the opportunities for using AI for health. This has been released in the public interest for all countries:

Guidelines From the WHO for AI usage

Protecting human autonomy

To protect the autonomy of humans in the context of health means, humans should be in control of the healthcare systems and medical decisions. Whatever information is accessible, it should remain confidential and be protected. Lastly, patients must give consent that is not only informed but valid as well through legal frameworks for data protection.

Supporting and promoting human well-being and safety

The regulatory requirements for safety, accuracy, and efficacy must be adhered to by AI experts for well-defined use cases. They must also keep a check on the measures taken for quality control. This is implemented to improve quality  in the use of AI.

Ensuring transparency and intelligibility

What this means is for ensure that all information is published or documented before the design or deployment of an AI technology. The information that has been updated must be easily accessible and also must integrate meaningful public consultation and debate on how the technology is designed and whether or not it should be used.

Fostering responsibility and accountability

Since AI are usually used to perform specific tasks, it is the responsibility of inventors to ensure that they are being used under contented and appropriate conditions, by appropriately trained people. Self checks, mechanisms and supervisions must be implemented by individuals and groups that make decisions based on algorithms.

Ensuring inclusiveness and equity

AI that is being developed for health must be designed to ensure the reach of the same is the widest possible equitable use and access. This is  irrespective of age, sex, gender, income, race, ethnicity, sexual orientation, ability, or other characteristics guarded under  human rights codes.

Promoting AI that is responsive and sustainable

Designers, developers, and users have transparency and readily available access to assess AI applications during actual use. Thus us done to determine how adequately and appropriately the AI responds to expectations and requirements. AI systems should also be designed to minimize their environmental consequences or be as environmentally friendly as possible. Governments and businesses should take control of anticipated disruptions in the workplace which includes training for healthcare workers to adapt to the use of AI systems.


These principles are aligned in place to guide the future WHO to support efforts to ensure that the full potential of AI for healthcare and public health is achieved to benefit us all. These reports state that the systems which are primarily trained on data collected from individuals in high-income countries would not be suitable for individuals in low- and middle-income settings.

Keeping those points in mind, AI systems should therefore be carefully designed according to socio-economic and health-care settings of that region. AI training models must implement community engagement and awareness-raising for millions of healthcare workers who will require digital literacy.  These guidelines are implemented to create a better environment for both workers and patients alike.

Want to learn more about AI in health?

NLP is being used to predict Covid-19 outbreaks

Healthcare is utilizing AI to detect breast cancer

How the health industry is adopting AI

Arrow Upward