As AI becomes more widely adopted, these new systems face both traditional cybersecurity challenges and new threats. For example, adversaries can poison the data used to train a machine learning model or subvert the input fed to an automated system to manipulate outcomes. These potential vulnerabilities pose substantial risks to consumers, since AI is increasingly integrated in products such as smart home devices, medical implants, autonomous vehicles, drones, and more, as well as IT systems used in sectors such as health care, education, and government. Addressing these challenges will require global cooperation. As transatlantic allies bound by common values and with a shared interest in promoting the digital economy, the United States, Canada, and the EU could take the lead by collaborating on a joint initiative for better cybersecurity for AI.
ITIF's Center for Data Innovation hosted a video webinar to discuss the transatlantic opportunities for promoting better cybersecurity in AI, and the ways in which policymakers could promote these initiatives.