AI CODE OF ETHICS
AI CODE OF ETHICS
SCOPE
The purpose of this ethics policy is to provide the basis on how Stampix, from an ethical perspective, builds solutions, and conducts business when it comes to our AI development.
We – at Stampix – strive to work with an ethical approach in developing our business and our services. We want to ensure trust into our AI systems, that the decisions made by our AI systems are impartial and ethical trade offs are balanced.
We endorse the UN’s Universal Declarations on Human Rights, the European Convention for the Protection of Human Rights and Fundamental Freedoms, the European Union Charter of Fundamental Rights, as well as the GDPR legislation. We also strive to meet the UN’s goals for sustainable development, as well as helping AI benefit the entire community, not just official commercial interests.
This Policy applies to everyone engaged in our services such as the board of directors, employees, consultants, customers, users and partners.
OUR PROMISES
- Fundamental requirements
To realise a trustworthy AI we make sure we abide the following fundamental requirements: Human agency and oversight, Technical robustness and safety, Privacy and data governance, Transparency, Diversity, non-discrimination and fairness, Societal and evironmental wellbeing en Accountability.
- Ethical principles
We ensure the ethical principles are taken into account when building, updating and executing the AI programs.
- Respect for human autonomy
Our AI systems are designed to augment, complement and empower human cognitive, social and cultural skills. We follow human-centric design principles and leave meaningful opportunity for human choice. This means securing a human oversight over work processes in our AI systems.
- Prevention of harm
Our AI operates in a safe and secure environments that are technically robust and ensure that they are not open for malicious use. Our AI systems are developed with a preventative approach to risks and in a manner that they reliably behave as intended while minimizing unintentional and unexpected harm, and preventing unacceptable harm.
All privacy and data is protected through our AI systems. The quality of the data sets used is paramount to the performance of our AI systems. All socially constructed biases, inaccuracies, errors and mistakes are addressed prior to training, in addition the integrity of the data will be ensured.
- Fairness
The development, deployment and use of our AI systems happen in a fair environment. We make sure our AI is free from unfair biases, discrimination and stigmatisation. We guarantee that our AI systems will never lead to people being deceived or unjustifiably impaired in their freedom of choice.
- Explicability
We strive to have transparency around our processes. The capabilities and purpose of AI systems are openly communicated and decisions – to the extent possible – are explainable to those directly and indirectly affected. In case a model generates a specific output or decision that is difficult to explain (such as ‘black box’ algorithms), we guarantee that the system as a whole respects the fundamental rights.
We will always work to comply with the Data Protection Regulations by openly and comprehensively explain:
- what the services we construct are for;
- how they can be used; and
- how we use and handle the data that we have access to.
While the above principles certainly offer guidance towards solutions, they remain abstract ethical prescriptions. In case tension establishes itself between the above principles, we will ensure that the overall benefits should substantially exceed the foreseeable individual risks.
- Responsibility and control
We will always clarify who is responsible for and has control over the service, as well as how the division looks between supplier and customer. By taking into account that unintended consequences can occur, risk management is of most importance to be able to protect valuable assets with appropriate security control. If such events occur, the events will be immediately analyzed and measures identified in order to limit the damage caused by the occurrence, while at the same time we strive to prevent similar situations from recurring in the future.
- Security and integrity
We must always ensure that there is a clear division of responsibilities for security functions because AI security is a part of our entire service responsibility. We shall also make sure that our customers and suppliers understand the risks and possible implications, and we shall, therefore, have a high level of expertise in AI security.