Toni Townes-Whitley argues that companies need clear ethics principles before crises hit. Her organization runs a free digital crime unit tracking botnets and sharing data with Interpol. She warns that AI systems fail on people of color, citing Joy Buolamwini's research, and says GDPR's new rights like algorithmic explanation are essential guardrails.
Chapters
You've got to have your principles in place before the disaster hits, by the way. You've got to set principles before the challenge comes. If you're starting to look for principles after a challenge, you're a little bit late. You have to set your principles of what kind of company you want to be and then you've got to implement them when the challenge comes. So we've got principles around privacy and security and transparency, showing our government customers their data in government transparency centers, cybersecurity centers around the world. I just launched two last year in Singapore and in Brasilia. This is about who we stand for and what we stand for. Protecting human rights, ensuring sustainability and ethics in AI. So what is that? Why is that important? It's important because what may not be in the news all the time is a few companies get the call every time there's a cyber attack. We get the call wherever the attack is, whether it be France, Belgium, the uk, the us we get the call and we start pulling the information that we have from a digital crime unit that we run that just tracks botnet activity around the world. We don't do this for a charge. We do this free of charge. We do this and we feed this information to Interpol, we feed it to the U.S. we feed it, we feed information on what's occurring in the cyberspace and this is the kind of challenge that we have to try to address. But not just cyber, beyond cyber. This AI conversation has been so amazing. We've been leading at AI, but as you lean into AI, you better have a point of view on privacy, you better have a point of view on the trade offs, you better have a point of view on things like algorithmic bias. Anybody heard of algorithmic bias? What happens when your facial recognition doesn't recognize faces of people who are brown or have different facial features, People that look like me. What happens because we say coding is neutral? Well, coding is not done by neutral agents. People code. And are people neutral? Not always. It doesn't even mean malicious. It just simply means not aware. And so some of the new work that's come on algorithmic bias. I love the work from Joy and I never get her name correct. Mwala Buini, I think just recently did some work on the coded gaze. Right. Some of you are nodding. Work that she's found on the bias, racial bias and algorithmic coding. Gdpr. A new privacy standard for Europe is coming. A privacy standard that says that you that every European has the right to understand how their data is used in an algorithm. They have the right to be forgotten. They have the right to audit. These are changing the way we start to build applications and how we deal with privacy and data. Because if you're going to step into AI and you're going to lead in AI, you better know what the end looks like. And the end looks like a conversation about how far are we going to go and what are we going to protect along the way.
