By Nadia Zaifulizan
The power of intelligence and autonomous technical systems are unequivocal. In line with its growth there are also increasing interests in establishing societal guidelines that contain intelligence systems to serve humanity’s values and ethical principles. In this entry we look at the goals of ethically aligned designs for these systems, based on the 5 guiding principles.
Human rights as defined by international law provides a basis of understanding that any creation must take into consideration the way that they affect humans, their emotions, their data or agency (prerogative). This means that human rights ideals within business, corporate, and public contexts can be adapted by engineers and technologists when creating any autonomous or intelligence systems (A/IS). This also applies in the way that the A/IS are operated, tested, and validated. A/IS must not infringe on human rights, freedom, dignity and privacy.
Autonomous or intelligence systems (A/IS) that are ethically aligned are able to advance the benefits for humanity by avoiding negative unintended consequences and increasing value for customers and society. This is often measured by parameters such as GDP, profit, consumption levels, but now it should also include a deeper evaluation. ‘Human satisfaction with life and the conditions of life as well as an appropriate balance between positive and negative effect’, also known as ‘well-being’, as defined by the IEEE Global Initiative, is an important measure that should also be considered by A/IS developers. Utilizing the best available well-being metrics as a reference point, A/IS should prioritize human well-being as an outcome in all system designs.
Issues of responsibility, culpability, liability, and accountability for autonomous or intelligence systems A/IS during the development and deployment stage must be clarified in order for manufacturers and users to understand their rights, obligations, and responsibilities. Designers, manufacturers, owners, and operators of A/IS are all responsible and accountable for the programming, output, purpose, application, and use of A/IS. Tracking systems for registration and record-keeping is important in order to determine the entity legally responsible for a particular A/IS.
Accountability is enhanced by transparency. Although the level of transparency will differ for each stakeholder, operation of
autonomous or intelligence systems (A/IS) must be transparent to enable user to understand how and why a system made a particular decision, or acted the way it did. Transparency also addresses the concepts of traceability, explicability, and interpretability, so that users can not only know the reasons behind its decision-making but also the capabilities and limitations of the AI systems that they use or interact with.
Technology Misuse and its Awareness
There is a greater risk of misuse the more powerful the autonomous or intelligence systems (A/IS) are. The impact of risks such as hacking, misuse of personal data, or exploitation require an all-encompassing participatory prevention. Designers and law enforcers have their own roles in guiding new technology towards positive outcomes and developing accountability structures that educate the masses. Raising public awareness on ethics and data security involves developers, government, lawmakers, enforcement agencies, across multiple levels of society.
What does This Mean?
The autonomous and intelligent systems (A/IS) today are instilled with increasing autonomy in making decisions and manipulating their environment. For the protection and preservation of human interests and the rest of society, it is essential that the A/IS are designed to adopt, learn, adapt, and abide by the norms and values of the community they serve. Transparency in executing compliance and ability to provide reasoning is also important. If humans were to develop some appropriate levels of trust in A/IS in the specific roles where the A/IS function, the transparency of their decision-making must be made clear.
The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems, Version 2. IEEE, 2017. http://standards. ieee.org/develop/indconn/ec/autonomous_ systems.html.