Governance and responsible AI

- [Presenter] In this one, we're going to talk about governance and responsibility. In particular, we'll examine these topics when it comes to developing AI-driven tools. Now, when we talk about AI, it's important to keep in mind compliances. You need to know the restrictions and regulations that are specific to the markets that you're operating in. It's also important to stay informed. This is because AI is a fast moving space and new regulations show up more frequently than perhaps what you're used to. Finally, you'll want to check your supply chain. Are you using third party APIs? Do they adhere to the same standards and compliances that you need to adhere to? Now, an interesting example of fairly recent AI-related regulation is the EU AI Act. Now, what's interesting about this legislation is that it lays out this framework for approaching risk mitigation in AI. So we have this pyramid of unacceptable risk, high risk, limited risk, and minimal risk. Now, while this legislation is of course more relevant to the European Union, it's very likely that other governments and legislative bodies will adopt similar legislation and regulation. Now, regardless of legislation, there are a few ethical considerations and practices you'll likely want to follow when rolling out AI-powered tools and apps. It's important to be transparent with users about their data, about how you train your models, and you want to seek explicit permission when using user data for perhaps training or improving models. It's also important to practice caution when dealing with sensitive applications in sensitive industries. Examples of such industries include finance and health, things that can really impact people's lives. So, while companies and startups want to adopt technologies as fast as possible, it's important to both stay informed of existing regulation and practice respect for our users and their privacy.