Interview with Karina Gibert – Head of the Intelligent Data Science and Artificial Intelligence Research Center (IDEAI)

• Artificial intelligence has become one of the hottest topics of the moment. What has led us to this?

The irruption of deep learning techniques where the first open door to scale up the AI algorithms with BigData and be able to solve complex problems in a really short computational time. Then, understanding that the intensive use of AI to extract knowledge from an organization’s data can provide a lot of added value at strategic and business levels promoted a big interest in the industry to introduce AI in their business processes. Privacy, biasses, and Explainability became then a challenge in the field, and very recently generative AI (an evolution of deep learning) has been making a lot of noise in the media and increased a lot the curiosity of the general population about AI and the interest of different economic sectors to take advantage of it.

• Are we about to undergo through a big innovation phase that will completely alter our current systems?

Not sure. There are still lots of challenges to be solved to see a change of paradigm. And one of the most critical is the carbon print of big data and AI algorithms.

• Within the automotive sector, can AI help us improve the field? How?

There are many AI applications that already did, like connected vehicles and safety improvement or AI-assisted industrial design of car components for example. IoT is for sure one of the driving technologies to develop more intelligent processes to predict vehicle failures in advance, to improve the design of components, to recommend the better vehicle to a specific type of driver, to improve production processes, to assist innovation in the sector or to help better driving according to weather or traffic conditions, among many others

• What ethical implications can be a barrier when using artificial intelligence to innovate in mobility? What current mobility challenges could be solved through the use of data?

Using the personal data of drivers or customers to train the AI models requires some caution in terms of privacy. But also, data emitted by vehicles themselves have to be treated with care, since it can provide information about the drivers which should not be disclosed, or should not be used by the company or governments. Also, for those autonomous systems in automotive cars, biases on decisions should be carefully under control to avoid undesired reactions of the system. The ethical implications of an autonomous car deciding to stop or not in front of a person crossing the street is a famous ethical dilemma that still is related to an open problem. Finally, cybersecurity vulnerabilities are a very critical challenge that requires attention.

• What can we expect in the short term from the development of new AI tools that affect the mobility sector? 

Most of the current novelties are well known, like advances in autonomous vehicles and intelligent management of traffic or public transportation for example. Advances in multichannel commodity delivery through the use of AI are impacting even the supply chain. But probably one of the most challenging fields that presumably can change the mobility sector in a very significant way is to see how last-mile delivery of small commodities can be covered with drones instead of terrestrial transportation vehicles. If this can be both technically, safely, and ethically solved the impact on traffic in urban areas will probably change significantly the behavior of cities.

• Can AI be developed to help us overcome gender-balance barriers and promote a more diverse and inclusive society?

There are many groups of women intensively working towards these goals. We need more women in the AI working teams, who provide the female view of the problems and the solutions, and who constitute referents for new generations. The lack of women in the sector has severe implications for the equity of society. From biased datasets that promote better men in regard to women to biased algorithms that deprioritize women systematically (when searching for jobs, when searching funds for their companies, etc etc). For years women got more damage in traffic incidents, just because the cars were optimized to be safe using male manikins for the testing. We still have a long work to do to attract more women to specialize in this sector, to see them working as professionals in the sector, and to scale up to directive positions to perform significant changes regarding the gender gap.