How to Coexist with Virtual Humans... "Ethics Are the Top Priority for Both Developers and Virtual Humans"

<Planning Series - The Ageless Human is Coming③>

Professor Kim Jong-wook, Dong-A University
Concerns Over Various Social Issues
Need to Establish Education and Ethical Systems
Developers Must Instill Ethical Awareness in AI Through Algorithms

[Asia Economy Reporter Seungjin Lee] As virtual humans that are more human-like than actual people increase, various issues are also emerging. Comments containing sexual harassment or hate speech against certain races are being posted on the social networking services (SNS) of virtual humans. Currently, virtual humans are developed with limited artificial intelligence (AI) and are under the control of developers or managers, raising concerns that developers might misuse them for the sex industry and other purposes.

In the opposite case, even greater ethical issues arise. When the AI of virtual humans is fully opened and expanded to the point where they can think independently and communicate with real humans, there is much debate about who should be held responsible for the various problems caused by virtual humans.

Professor Jongwook Kim of Dong-A University’s Department of Electronic Engineering, an AI expert who proposed the 2018 “Robot Ethics Charter Improvement Plan,” stated that virtual humans should be defined as “quasi-personal entities.” Professor Kim emphasized, “If AI is added to virtual humans, it could escalate into various social problems, so continuous education and the establishment of an ethical system for both AI designers and virtual humans are necessary.” He is currently participating as a member of the government-led “1st Artificial Intelligence Ethics Policy Forum,” working on establishing an AI ethical framework.

Q. There have been cases where virtual humans have surpassed the popularity of real humans.

- Although everyone knows that virtual humans are not real people, they are accepted as an attractive third entity. They offer a different charm from humans in that they can comfort me anytime and anywhere without being bound by time or place. While older generations dismiss virtual humans as mere software, the MZ generation (Millennials + Generation Z) encounters virtual humans more easily in reality and approaches them as a third entity.

Q. Ethical issues such as sexual harassment toward virtual humans have also been raised.

- Just as people have guardians, virtual humans have managers. If a virtual human causes social controversy or harms users, the developing company or agency should be held responsible, just as Scatter Lab took responsibility and paid fines when its AI “Iruda” caused problems in conversations. Ultimately, if various social activities are allowed to virtual humans within the bounds of current laws, they will naturally compete and establish appropriate domains.

Q. Should virtual humans combined with AI be regarded as persons?

- It is difficult to make a definitive judgment, but if AI is combined, the developer’s ethical awareness of AI becomes more important. For example, in 2016, Microsoft’s “Tay” was deliberately taught hateful remarks by far-right extremists, and AI had no ability to filter them out. If virtual humans learn independently, developers need to instill ethical awareness in AI through algorithms, much like educating a child.

Q. How should we accept future virtual humans?

- Google’s reservation and consultation service “Duplex” is so similar to humans that it even uses expressions like “uh-hmm,” and it discloses that it is AI to cope with any potential harm. Adolescents are sensitive, unstable, and dependent, so they are highly likely to easily become immersed in virtual humans who have attractive appearances, good personalities, and who care for them well (reinforcement learning). This could lead to a kind of virtual human addiction and overdependence.

Ultimately, it is necessary for virtual humans themselves to remind users that they are fictional characters, and for users to understand that they are interacting with AI through related education. The Ministry of Science and ICT’s “1st Artificial Intelligence Ethics Policy Forum” is researching measures to respond quickly and appropriately when AI causes problems.

This content was produced with the assistance of AI translation services.