How Do You Want to See AI Used?

AI makes hundreds of micro-decisions for us daily. Whether it's analyzing your shopping habits to give you recommendations or sweeping social media sites to remove offensive comments, artificial intelligence drives online life. And it has the potential to do much more.

In a PCMag survey on future technology, 32 percent of respondents wanted AI assistance with early detection and prevention of medical problems—something AI has already tackled successfully>. Fighting climate change and allocating resources for the impoverished came next at 19 and 17 percent, respectively. Detecting cybersecurity threats (where AI is most prevalent) came in at 11 percent. Justice reform and accuracy tied with disaster response at 9 percent.

AI is already used in many of those fields and more besides. Unfortunately, it can't always be trusted.

Systems intended to be autonomous and impartial develop their own biases: That's a side effect of deep learning. If you train an AI on thousands of pictures, videos, and documents, it's going to pick up the prejudices already present in that information pool. Some biases are innocuous, others unintentionally discriminatory—and if you're tasking AI with high-stakes challenges such as social justice and medical intervention, bias could be a major problem.

Before we can feel comfortable ceding authority to AI, we need to understand how and why artificial intelligence acts the way it does—and given that PCMag's survey also revealed 27 percent of respondents were "not at all familiar" with AI, we have a long way to go.

Recommended by Our Editors

AT&T Shape: Where AI Meets Next-Gen Immersive Storytelling

Facebook's AI Is Good at Detecting Spam but Struggles With Hate Speech and Harassment

What Is Artificial Intelligence (AI)?

AI tech can identify genetic disorders from a person's face