On May 26th, 2023, the CMU Portugal Program hosted the Talk “Human-Centered and Responsible AI” by Carol Smith, a senior research scientist in human-machine interaction in the Carnegie Mellon Software Engineering Institute’s Artificial Intelligence Division, which addresses the need for leap-ahead AI capabilities that are reliable, responsible, safe, fair, and transparent.
Under an invitation by the CMU Portugal Program, Carol Smith spent a fruitful morning learning more about some of the initiatives in the field of Human Computer Interaction (HCI) and Responsible AI in which the Program is involved.
Nuno Nunes, CMU Portugal National Director and ITI President, started by sharing an overview of the Program and its main activities.
Shuhao Ma and Pedro Ferreira had the opportunity to present their work under Ph.Ds. Shuhao Ma is a CMU Portugal Affiliated Ph.D. student in HCI and an HCI design researcher at ITI, LARSyS. Pedro Ferreira is a CMU Portugal Affiliated Ph.D. student whose primary interest is research and development in Child-Computer Interaction and the participation of educational communities. The discussion also included the participation of Duarte Ferreira, a CMU Portugal affiliated Ph.D. Student.
The presentations continued with Hugo Nicolau, Assistant Professor at Instituto Superior Técnico and researcher at the Interactive Technologies Institute / LARSyS. Hugo presented the DCitizens project that aims to build the capacity to conduct research in the field of Digital Civics. Particularly how digital technologies can scaffold a move from transactional to relational service models and the potential of such models to reconfigure power relations between citizens, communities, and institutions.
João Guerreiro, Assistant Professor at Faculdade de Ciências da Universidade de Lisboa and a researcher at LASIGE, followed the conversation with the highlights of HCI being conducted at this institution mainly in Human-Computer Interaction, Non-Visual Interaction, and Accessible Computing.
Helena Moniz, President of the European Association for Machine Translation, Vice President of the International Association for Machine Translation, and Vice-Coordinator of the Human Language Technologies Lab at INESC-ID presented next the new project Center for Responsible AI, where she Chairs the Ethics Committee. This Center, led by the Portuguese startup Unbabel and launched within the Portuguese Recovery and Resilience Plan, is one of the largest centers dedicated to Responsible AI.
The discussion sessions concluded with the CMU Portugal Talk “Human-Centered and Responsible AI”. The event gathered over 30 participants at CMU Portugal/ITI Premises at Factory (Hub Criativo do Beato). During this talk, Carol Smith focused on new insights for building AI systems that are human-centered and responsible to support participants in creating systems that people are willing to be accountable for.
According to Smith, through decades of HCI and UX efforts, excellent methods and practices have been developed. Still, the dynamic nature of AI systems requires more speculative thinking and more flexible designs. For example, there’s the need to decide which decisions systems are allowed to make; what exchanges between humans and machines will occur; how the interactions will change over time; and how people can be kept safe and in control. It is also crucial to focus on problems such as: how the system will share information on which decisions were based on, how it will convey its capabilities within new contexts, and how the system’s confidence in its responses will be shared.
The talk ended with a Q&A by the audience focusing on some issues related to the importance of Responsible AI, ethics, and the role of human-centered design, namely in serving but also preserving humans and a younger generation of users. The discussion continued during a light lunch at the premises, an opportunity to network and open doors for future collaborations.