Content - Session VII: AI Applications in Security, Safety and Communication

AI Applications in Security and Safety

Session abstract

One of the best-known uses of artificial intelligence is in image and, more recently, facial recognition. This is also an area where AI has shown tremendous progress in recent years. This technology is already employed in many systems designed to ensure security, for example in criminal surveillance, but also on an individual scale or for information protection. Indeed, when combined with biometrics, AI allows for more secure authentication, and by extension, data protection. However, AI is not only about security. It can also bring greater safety in our daily lives, a concept that is often wrongly confused with security. Digital transactions have reached unprecedented levels and continue to grow, yet the digital world is complex and holds its share of dangers. Too often, users are vulnerable to cyber-attacks and to being duped by fraudulent products or transactions in the digital and physical worlds. AI has the potential to make a difference by detecting fraud and protecting systems from cyber-attack. Despite the many uses of AI in security and safety, there are concerns. Experts warn of AI repeating human biases like discrimination.
This session aims to show how AI can be used to create a safe and secure environment in the digital and physical worlds while remaining a strictly supervised tool unaltered by human preconceptions. To do so, this session will draw on experts from leading developers of safety and security technologies using AI, as well as promising and innovative newcomers addressing lesser known but equally critical issues.

Research & Development of Face Recognition Technology in NEC
Chair: Ken Hanazawa

This presentation will explain what facial authentication is and its advantages among other biometric technologies. It will also give an overview of NEC activities and strengths in facial authentication technologies and their applications.

Defensive and offensive AI for embedded cyber-security
Chair: Yan-Taro Clochard

Artificial Intelligence - A Gift & a Curse with Respect to Security & Privacy
Dr. Michael Tagscherer

We and „things“ around us become more and more connected. Hence, the internet of things (IoT) will change our lives dramatically. Applications will be “smart”, based on data gathered from connected devices and us! This growing amount of data can only be processed and analyzed by applying Artificial Intelligence (AI). Being a powerful tool, AI can also be used to manipulate data. e.g., “mimic” the voices or videos of real persons. Because of AI, it is becoming more difficult to judge if information is trustworthy (e.g., integrity, origin). This is an issue for us, but also for any kind of data driven application. Some light shall be shed on the opportunities – but also on the threat side – of AI.

Artificial Intelligence: From Research to Business Value
Virginie Haas

AI-induced Discrimination: Risks & Solutions
Anna Choury

AI is a powerful tool for improving processes in almost every aspect of modern society. If not trained responsibly, however, there is a risk of AI being unfair. Scandals have appeared in recent years involving AI-induced racism, sexism and homophobia. Machines are reproducing society's biases, but this does not have to be the case. This presentation will discuss the pitfalls of AI and the possible solutions to ensure its ethicality.