Ethical principles
We apply eight ethical principles to the development and use of AI:
Human-centricity
As a public sector organisation, we serve people and communities. That is why we develop services that utilise data and AI for the benefit and with the input of people.
Transparency
People should be able to access information about the use of data and AI in an understandable format, so that they can evaluate the City’s activities. The only way we can build trust is through openness. That is why we communicate as clearly as possible how and where we use data and AI. We actively seek to share all relevant information, unless prohibited by law. We highlight any identified shortcomings in data. Where possible, we also share the source code.
Explainability
If there are significant risks associated with the use of data and AI, we are able to explain a specific result or the general operating logic of the algorithm in a way that is easy to understand. Explainability helps us to manage the system’s impacts. If the operation of AI has legal implications, explainability helps us to ensure that the customer’s rights are upheld.
Fairness and equality
Our use of data and AI solutions is based on respect for the dignity and rights of every individual. The use of data and AI systems can replicate and reinforce existing discriminatory structures. That is why we take account of any potential biases in data and algorithms and address them where necessary. Where necessary, we apply positive discrimination to promote equality and non-discrimination.
Responsibility and trust
We want our customers to be able to trust our operations when we use data and AI. We designate a responsible party whom our customers can contact for every service that uses AI. The responsible party ensures that the ethical principles governing data and AI are adhered to.
Privacy
We process personal data with care and in a secure manner throughout the entire lifecycle of the data. By safeguarding privacy, we can make use of data securely and reliably.
Human control
In some situations, managing the risks associated with a system requires human control. When this is the case, we ensure that the person responsible is able to monitor and supervise the operation of the system and, where necessary, intervene in it. An AI system cannot learn every possible scenario, and there is always a risk that the system will make an erroneous conclusion or encounter a technical error. When the system is under human control, we are able to respond to unexpected situations and impacts.
Security
We ensure that the systems that use data and AI are effectively protected and managed. We identify and minimise potential risks in data and its processing, as well as in AI systems and their use. Risk management may restrict our ability to share information about systems. We minimise potential security risks so that city residents can benefit from the use of data and AI.
Environmental impacts
The development and use of AI consume energy and other resources.
Helsinki assesses the environmental impacts of AI solutions before they are deployed and also monitors their impacts during their use.
Our aim is to use AI in a way that supports the City’s climate goals and helps to develop services that are both effective and environmentally sustainable.