How tenuous is the position of automated stock trading algorithms when we see that all it takes to confuse a state-of-the-art self-driving car is some table salt? How reliably can the military trust the self-navigation of AI-powered drones when AI vision has been fooled with only carefully placed stickers? Will you take stock in an AI doctor's diagnosis if you know that cutting edge healthcare algorithms miss a cancerous lesion when its photograph is slightly tilted?
AI systems have become a key technology to automate government, fuel businesses, and arbitrate everything from your credit score to your exercise habits. But, since they are left unguarded, they are prime targets for adversaries. Pioneers in machine learning -- Google, Tesla, Microsoft, Amazon, OpenAI, Facebook, IBM, Tumblr, YouTube - have already had their algorithms tricked, stolen, and evaded by adversaries.
AI is brittle and you are about to see how attackers are breaking it for their fun and profit.