“Today’s challenges call for human-centered approaches”

Head of the Social Computing research group, Daniel Gatica-Perez has contributed to Idiap’s vision of Artificial Intelligence for Society since 2002. Last year, he received two awards recognizing his long-term impact at the interface between technology and society.

Today, smartphones are standard scientific tools, used to collect and analyze real-world data. Back in 2002, when Daniel Gatica-Perez joined Idiap Research Institute after a PhD in the United States, it was still five years before the release of the first iPhone. His career at Idiap is intertwined with the digital and mobile revolutions that have transformed our societies. Gatica-Perez is also Professor at EPFL since 2014, and the director of the Digital Humanities master’s program since 2018. Some of his publications are among the most cited in his field. We met him for an interview to reflect on our long-term relationship as societies with ubiquitous and mobile technology.

Why did you choose to join Idiap in the early 2000s?

I joined the institute in the framework of the Swiss National Center of Competence in Research on Interactive Multimodal Information Management, also called IM2. It was a crucial project for Idiap’s development. IM2 allowed Idiap to rally around an integrative vision: a multi-sensor meeting room that brought together technologies like speech processing, computer vision, and text mining. At the time, I was interested in studying small groups working together and supported by technology. This work was an opportunity to mature perspectives related to how to approach research involving people and technology from a human-centered perspective.

From this point of view, how did mobile devices change your work?

The early days of smartphones were an opportunity to go outside the lab and conduct studies in the field. Understanding how phones are used in everyday life is important. Ten years ago, in collaboration with Nokia, we were able to curate and share an open mobile dataset, used over the years by hundreds of researchers worldwide for academic purposes. This topic is still relevant today. As one example, much research in the domains of health and wellbeing is currently done using smartphones and mobile apps as platforms.

This opportunity also created risks in terms of privacy, didn’t it?

You’re right, privacy is a fundamental issue in human-centered computing, but today, research goes beyond privacy. Another key issue is diversity. When technology only serves the interests and reflects the views of certain groups in society, there are risks of reproducing and deepening existing divides. Some of these risks can be mitigated by designing technology with and for people. Smartphones are valuable tools to conduct participatory research based on citizen action.

Mobile and social technologies have given people a voice and allow us to engage in themes of collective value, be it community health or urban concerns. Especially when working outside the Global North, one must go beyond the current AI hype, and deal with the fact that locally contextualized and valid data used to train machine learning models is limited. For example, over half of the data in the most popular image dataset used to train visual recognition systems comes from only two countries (the US and the UK.) We need to systematically increase diversity in datasets.

Big tech companies play a major role in this field, is there room for public research to shape these technologies in a more inclusive way?

Yes, I believe that we, as researchers, have a role to play. Academic research cannot always compete with big tech companies that, just to give one example, have almost unlimited computing resources. On the other hand, research in computing with a societal focus has its own agenda. We have to acknowledge that for grand challenges like public health or climate change, quick fixes do not exist and solutions will not come from technology only, and that there is a need to understand both societal conditions and individual human experiences, and, ultimately, that we have to involve people in such processes.

We should also aim for multidisciplinary, human-centered approaches. For example, in the framework of the SenseCityVity project, launched in the mid-2010s, we worked with local partners in Mexico and designed an urban challenge, inviting participants to use their smartphone to collect multimedia data to map and document urban issues deemed relevant by citizens. Later on, thanks to its success, the project was extended to other countries. In Switzerland, we created the “civique” mobile platform that has enabled several local projects, more recently the Corona Citizen Science project, where people shared their experience during the 2020 covid pandemic in Switzerland. Participatory, multi-disciplinary science is needed for all these challenges, and human-centered computing can contribute to it.

More information

- Social computing research group at Idiap
- SenseCityVity research project
- Corona Citizen Science project
- ACM ICMI Sustained Accomplishment Award
- ACM Ubicomp 10-Year Impact Award