Making sense of the roles within design of ethical AI: designer’s and organisation’s responsibility, and a matter of public discourse?
Introduction
This writing will focus on discussing the roles of the designers, organisations and the open public discourse in the context of designing AI systems in an ethical manner. The central question presented in this essay is that what role should these agents assume in the context of responsibility and participation over the design of the AI systems, and why?
Three different AI ethics principle frameworks suggesting different solutions will also be briefly discussed; 1) The Future Computed: Artificial Intelligence and its Role in Society [1], 2) Chinese Governments AI Alliance draft [2] and 3) Tenets from AI Partnership [3]. Based on the insights from these frameworks, throughout the essay it is argued that:
(1) successful participation of the publics requires effective education and sharing,
(2) that design of organisations and organisation cultures is a central constituent of responsible AI design,
and that
(3) the designers should take on the role to learn from, develop and utilise the existing knowledge as well as to share their process openly
Co-constructing the future of AI from the organizations’ perspective
The designers, organisations and people co-construct the society and the conditions of the current reality. Based on this constructivist viewpoint it becomes relevant to discuss, how much each of these agents should assume responsibility and what could be their role and degree of participation in forming the emerging AI systems?
Firstly, in order to participate one needs to understand the matter of discourse and concern. It has been highlighted by [2] and [1] both, that open discussion and sharing with the public is a required action in designing AI in an ethical manner. This means, that organisations or academia — the executing agents or organisations — can be assumed to have this responsibility for sharing the knowledge and educating the publics. The next question is then, how should this sharing take place, should it be direct or indirect, and how should sharing of various aspects be conducted?
One viable pathway for this sharing is open source procedure and publishing the content and the projects in a transparent manner online, where all members of the society have access to the information they wish to access. Additionally, the individuals will need a certain amount of “AI literacy” in order to be able to understand, critically assess and discuss on contribution the subject matter. The level of AI literacy determines, how advanced the content of writing should be; and therefore it would be reasonable to utilise different pathways of communication targeted to different kinds of people with various levels of AI literacy. For designers specifically (as central agents making decisions over the designed technology), these should be shared on platforms, educational courses, and other venues that designers commonly are exposed to. The knowledge-sharing may even take place in organizations’ internal training, but it would be important to keep the knowledge open for others to access: outside of the organization and together build the knowledge base and discussion landscapes for ethical AI. Closing doors and “withholding the information for competitive advantage” is lastly needed, and in this way the ethical goals might have a contradiction with purely capitalistic gains goals.
As mentioned above, organisations and institutions have the central role of executing AI development. In these contexts, it seems that in a general level of sharing and communication within the organization is vital, as mentioned by [1][2][3]. More specific points that are raised up are strengthening the diversity of the people working at these organisations and developing the AI systems, advancing the awareness of the social responsibility [2] and creating a culture of trust and co-operation [3]. Therefore, it can be concluded that the design of the organisation and the organisation culture become highlighted and are central constituents on the responsible design of AI.
Complexity of creating change
Another notion with the aim of participation of this wide of a scale is, that it is more complex and traditionally humans have failed in it to a certain extent. One can not argue that disadvantaged voices from developing countries would have the same power to contribute as the employees of a tech company in a western country. We all have individual levels of agency — but nevertheless this should be the goal glimmering in the horizon. Collaboration can take more time, the more contributing agents, with the assumption that the goal is for everyones voices to be heard. One way to address this issue would be to intentionally slow down the technology-driven development through regulations set for companies and academic research. However, it is potential that this kind of regulation would effectively hinder innovation in the domain and would therefore be controversial approach. I would therefore suggest as a moderate liberal approach of proactively educating the people as effectively as possible and ideally making sure, that all members of the participating society keep up with the pace of the development as well as possible. [2] has also similarly highlighted the importance of educating the public in regards of AI ethics and the design of the systems. In the end — change is always going to take place and we need to learn to adapt to it and work with it as a society. Maybe the goal is, rather, to understand how we can establish effective knowledge-sharing flows and platforms that allow the knowledge to spread from top to the bottom fast — and this way work towards equality and inclusion.
Designers perspective
Following the requirements for two of the above, some conclusions can be derived for what the individual designers role should be. In order to make informed decisions, the designers should constantly aspire on learning and keeping up to date with the development of AI in order to design as responsibly as possible. The designers also should practice the above discussed culture of sharing, co-operating, both within the organization as well as with the whole society and proactively use the agency they have. Channels such as Medium can be used effectively in this for AI designers discuss their design process and the issues they have faced, of which other designers can learn and reflect on — and grassroot organizations and communities can be established for discussion platforms. This is something that has a long tradition in design sphere and that designers are rather good at: establishing venues of sharing their work, discussing, and learning from each other.
Conclusions
Designers, organisations and participating people in the society co-construct the conditions of the current reality: and their actions have an effect on the emerging futures to varying extents. An approach for ethical design of AI systems is to diversify the participation and focus on including all of these different agents of change into the development process— and to assess the best strategies and approaches for doing this.
The challenge that we all should be thought about is how should the sharing, education, learning, take place on a detailed and practical level — and this is one relevant matter of discourse to address in the future. But ultimately, we need to develop a high level of participation through open sharing and discussions, and prime our communities for the emerging AI futures. Therefore — as many have already argued — that people should be the foremost focus in developing ethical AI futures. And we need to address the prevailing societal inequalities in order to achieve these goals.
References
- The Future Computed: Artificial Intelligence and its Role in Society. Retrieved December 7, 2019 from https://3er1viui9wo30pkxh1v2nh4w- wpengine.netdna-ssl.com/wp-content/uploads/2018/02/The-Future- Computed_2.8.18.pdf
- Chinese Governments AI Alliance draft. Retrieved December 7, 2019 from https://www.newamerica.org/cybersecurity-initiative/digichina/blog/translation-chinese-ai-alliance-drafts-self-discipline-joint-pledge/
- AI Partnership. Tenets. Retrieved December 7, 2019 from https://www.partnershiponai.org/tenets/
