Ethical dilemmas of autonomous machines. Sii expert talks about challenges in designing artificial intelligence
Challenges faced by the manufacturers of autonomous machines, relationship between ethics and technology, and the future in the field of user experience design are just a few of the topics raised by a Sii expert during his speech at the EuroIA conference in Dublin.
The EuroIA conference is one of the most important events in Europe dedicated to information architecture and user experience design. This year, the 14th edition of the conference took place in Dublin, and the speakers included Sławomir Molenda, UX Designer from the Sii Digital Competency Center. After his return, he told us about the most important aspects of designing ethical algorithms, which are used in the production of autonomous machines, as well as about the prospects they open for specialists designing user experience and user interface.
Barbara Kun, Marketing Account Manager: Your presentation was entitled The design of ethical algorithms in autonomous machines. What did you want to emphasize in your speech?
Sławomir Molenda, UX Designer: Designing good (as opposed to harmful) artificial intelligence is one of the challenges of the upcoming fourth industrial revolution. In my presentation, I wanted to describe the related technological and philosophical problems, e.g. how to ‘embed’ ethics in autonomous machines, or what ethical norms they should be guided by without harming us at the same time.
Many techno-optimists believe that greater automation will ultimately free mankind from tiring and repetitive activities, that we will spend time on pleasure alone. This builds a narrative of progress quite well.
However, thanks to advanced artificial intelligence, we want to get rid of responsibility when making difficult decisions. Responsibility would rest on algorithms in cars that do not require drivers, exploratory and rescue robots, military systems (e.g. drones), medical robots, but also sex robots that will be gaining more autonomy.
Modern algorithms are still being written by people. It is easy to make various kinds of cognitive mistakes in them, or not to predict all scenarios. We should also ask ourselves whether we are technologically and ethically ready enough to turn over the decisions concerning someone’s life or death to machines.
B.K.: In your presentation, you gave examples of ethical dilemmas known from philosophy, such as a trolley whose change of track will save five people at the expense of one. It seems that issues which so far have been only a pretext for theoretical considerations can find quite practical application. How can ethics be combined with technology?
S.M.: Since the eighteenth century or so, we have been experiencing huge social changes thanks to the development of industry. Ethics evolves together with us, and social norms change. We can see that technological progress also has many dark sides. It can alienate us, dehumanize us and lead to greater nihilism.
Together with the development of artificial intelligence, the importance of ethics in AI design has grown significantly. It has been added to the agenda of the largest IT corporations and EU institutions. I sincerely hope that the result will be a positive one.
There are several problems related to the design of ethical artificial intelligence. The first one is poorly prepared data for algorithms training. As a result, the outcome of their work can be unfair. In her book Weapons of Maths Destruction, Cathy O’Neil gave many examples of poorly designed systems. They are able to discriminate different groups on the basis of gender and skin colour, treat harmless prisoners in the conditions of recidivism, and create enclaves for privileged people such as white rich men.
There are also products driven by artificial intelligence, which, by definition, have been designed for unethical activities. I am thinking here of systems supporting manipulation of people, spreading disinformation and fear.
But it is only the beginning of the development of strong artificial intelligence and the ethical problems that it will bring. In the near future, we will see the consequences of decisions taken by autonomous cars. In crisis situations, they will have to choose between the life of a driver or, for example, a group of people at a pedestrian crossing. We must remember that when we talk about algorithms, we are also talking about their developers or guidelines they followed.
Such ethical considerations provoke and force us to think about what ethics really is. Are there any universal norms that guide people? Is it at all possible to transfer human reasoning and ethical context to autonomous machines? These are very difficult questions for philosophers and developers alike.
B.K.: Sławomir, why did you decide to deliver this presentation during the EuroAU conference?
S.M.: EuroIA is the only conference in Europe where information architecture issues are really important for its participants and organizers. For example, this year I was able to take part in the ‘Ontology Dojo’ workshop. We practiced there using ontological thinking to create objects, their predicates, attributes and classes for these objects.
Apart from interesting workshops, EuroIA is visited by very open people. The atmosphere is friendly and family-like. You can easily talk to the most experienced practitioners from the design industry, and also well-known authors of books on information architecture, such as Donna Spencer, Peter Morville or Louis Rosenfeld.
But despite this, the conference is still intimate and remains in the shadow of huge events such as UXLX or UX London. Probably because it has no ‘UX’ in the name.
B.K.: What is the impact of changes in the IT industry on user experience and user interface design?
S.M.: I don’t know if this is due to changes in IT or better cooperation of designers with developers and other designers. But we can certainly create prototypes faster. It is supported by better tools, their plugins, UI control libraries, various design systems (Google, IBM) and applications for interface animation.
However, the phases preceding the creation of the prototype still take a lot of time, regardless of changes in IT. Ideation, discovering potential solutions (e.g. thanks to Design Thinking), ethnographic research, collecting business requirements. These are the stages that often in fact determine the success of a product.
B.K.: You are an expert with over ten years of experience. You have had an opportunity to work on many commercial projects for Polish and foreign companies. In your opinion, what prospects are opening up for specialists in the area of user experience and user interface?
S.M.: Companies cooperating with Sii, which care about the satisfaction of their customers, employees and better products, ask directly for the support of UX designers. I believe that with such intensive development of the IT industry and demand for better design, there will be even more work for us.
At Sii, I have an opportunity to work for customers representing various industries – automotive, financial, medical, construction or sports ones. This diversity means that I do not close myself in one, specific domain of knowledge. It is really developing.
Apart from that, each of these industries is undergoing changes related to new technologies. There is a growing need for specialists in voice and conversation interfaces, virtual reality, augmented reality and artificial intelligence behaviour. Prospects for the UX industry are very interesting, and certainly nobody will have time for boredom.
We encourage you to get acquainted with the presentation entitled The design of ethical algorithms in autonomous machines available at slideshare.net
The Digital Competency Center specializes in designing User Experience as well as in software development, and construction and maintenance of IT systems. It supports companies in creating mobile and web applications, e-commerce solutions, managing business processes and implementing Robotic Process Automation. The complete offer can be found at the Center website. Visit us!