Privacy, Security and Ethics: The Challenges of AI for the Computer Scientists of tomorrow according to Dean Marc Langheinrich

Much of the current affairs concerning the digital sphere touches upon the areas of interest of Prof. Marc Langheinrich, Dean of the Faculty of Informatics at Università della Svizzera italiana (USI), who explored these issues of public interest in an interview published by Corriere del Ticino.

The development of new technologies, as explained by Professor Marc Langheinrich, will introduce new challenges for cybersecurity: "Security will be significantly impacted by generative Artificial Intelligence (AI), such as ChatGPT and similar tools. These will enable the ‘scam industry’—which is already as large as the entire illegal drug trade—to refine and significantly increase their activities with deceptive emails, fraudulent chats, and ever more convincing phone scams, including personalised voices."

Privacy protection will also become more difficult due to these new technologies: "The increase in wearable devices that collect physiological data—such as smartwatches—will make it increasingly hard for users to control some of their sensitive data. We have already seen an example of the implications, when the well-known genetic analysis start-up 23andme.com went bankrupt and now millions of the most intimate DNA records of their former clients seem to be available to anyone willing to buy these leftover assets, with no regard whatsoever for the original privacy policy, of course."

Dean Langheinrich’s research addresses the themes of security and privacy from a social and political perspective, because "although many security issues can indeed concern the individual first and foremost, privacy and security have always had a social significance. The more laptops and phones get hacked, the harder it becomes to keep the remaining ones secure," said the USI Professor.

Another key point is the societal value of privacy, as Professor Langheinrich illustrated with a concrete example: "While knowing the travel history of a single person may not be particularly important to a tech company, knowing the movements of over half of a country’s population is a huge asset that can be monetised in many ways. Switzerland must ensure that all this data is used for democratically legitimate purposes, that benefit its citizens, and that its citizens remain safe and are not exposed to online attacks from both outside and within its borders."

However, it is essential that users also play their part in protecting themselves: "The legal protection in place both in Switzerland and across the EU is suitable for safeguarding our privacy when it comes to the Internet of Things (IoT), Smart Homes and Wearables. However, users must also do their part by ensuring that the devices they purchase and install—or wear—comply with these principles. Not all devices found on cheap online marketplaces are a good choice in this regard. As users, we must pay attention to the security updates that will protect these devices from online threats. Unfortunately, current legislation is still unable to compel device manufacturers to provide these updates for a minimum number of years."

Given the challenges that privacy and ethics in computing present to today’s and tomorrow’s computer scientists, USI offers ethics and security-related courses; however, as Professor Langheinrich pointed out, the Faculty of Informatics intends to expand its course offerings in this area. "The university must certainly take a central role in strengthening the ethical competences of informatics graduates; more broadly, however, there needs to be a focus on developing the skills necessary for conscious use of technology from a much younger age."

The Dean also raised an additional issue: it is vital to ensure that all citizens remain integrated into society and can participate in daily life, regardless of whether they use new technologies. "Generative AI and voice interfaces could offer interesting ways to facilitate such participation for everyone, simplifying interactions and providing personalised guidance when needed," commented Professor Langheinrich.

Reflecting on AI’s potential, the Dean reminded us how this technology can be a valuable ally in various and practical fields: "Computer science offers solutions for global challenges such as climate change—for example, in the form of computational support for climate modelling or the development of more energy-efficient programming languages. The work on Explainable AI carried out here at USI aims to make AI models more transparent, thus enabling developers to clearly identify the limitations of these technologies, policymakers to ensure they are free from bias, and citizens to better understand how results are generated, providing them with tools to challenge potential errors. If we do not pay attention to biases and the lack of transparency in AI systems, we will face growing social inequality, since simplistic approaches to AI tend to amplify existing imbalances.” To prevent the technologies mentioned from becoming a source of new forms of inequality, strong legislation is required—legislation that ensures “data is not simply collected by large multinational corporations and monetised elsewhere, but that it benefits society and individuals equally,” concluded the Dean of the Faculty of Informatics.

The full interview with Professor Marc Langheinrich is available at the following link.