ADAPT Radio’s HumanAIse series continues this month with a discussion on how regulation, cyber hygiene and digital literacy are key to using AI for its fantastic benefits, while protecting us from the potential dangers it can expose us to.
AI systems are developing at such a rapid pace and have sparked questions about how we can control their use and prevent them destroying society. With the AI Act due later this year, we hear from two experts on how regulation, cyber hygiene and digital literacy are key to using AI for its fantastic benefits, while protecting us from the potential dangers it can expose us to.
Both our guests in this month’s podcast have extensive experience working with technology and safety, founder of the University of Pittsburgh Institute for Cyber Law, Policy, and Security, David J. Hickton and Head of the Artificial Intelligence Discipline at Trinity College Dublin’s School of Computer Science and Statistics, and interim director of the ADAPT Centre, Professor Dave Lewis.
Dave Lewis kicks off the podcast with a discussion on the potential risks versus the benefits of artificial intelligence. AI has the potential to impact various disciplines from medicine to law to computer science and even academics are starting to consider how it impacts their teaching practice. Areas such as medicine, software engineering and the natural sciences and the potential to be hugely benefited by AI, according to Lewis. However, one challenge is how we anticipate the harms and mitigate the risk of such technology which is where the EU AI Act can provide useful legislation to regulators.
David J. Hickton provides more information around the AI Act and how it has created a classification of various types of AI depending on the seriousness of the application and whether it requires more stringent and prompt regulation. Low risk applications, such as video games, require almost no regulation while higher risk applications, such as autonomous vehicles or Governmental applications, where there are potential consequences for society, need rigorous testing and accountability. Hickton also provides some background on how AI is affecting the US. He refers to how the legal community in the US is affected as AI perform routine tasks like cataloging documents, preparing standard court pleadings and submissions which could reduce staff and increase output. This becomes threatening to the private practice of law. He emphasizes the need to “give equal measure to the benefits as well as the concerns”.
Lewis continues the discussion with a brief look into why companies pushed back on the implementation of the AI Act back in 2021 when it was first proposed referencing qualms around the cost of such implementations. Hickton also describes the need for establishing international norms for digital space in response to misuse of technology while Lewis raises the need to quickly build up dialogue between the public, companies and regulators. As the AI Act may impact employment regulation, we need to figure out how to address it in high-risk areas such as education and democracy. Finally, Hickton raises the question of whether we will control this technology or will this technology control us?
For further insight into the AI Act, catch HumanAIse on SoundCloud, iTunes, Spotify, and Google Podcasts.
ADAPT Radio: HumanAIse is ADAPT’s newest podcast series providing an in-depth look at the future of AI, automation and the implications of entrusting machines with our most sensitive information and decisions.