Spark 434
Taking things apart, selling our own data, tricking AI to do the wrong thing, and a gender-inclusive stock pho
Gender-inclusive stock photo library goes beyond clichés
Stock photography isn't something most of us really think about. But those seemingly innocuous images we see —the photos that accompany news articles or ads — have the power to shape our perceptions. Spark host Nora Young speaks with Zackary Drucker, the photographer behind The Gender Spectrum Collection, a new stock photo database featuring trans and non-binary models.
Selling our own data could cut out middlemen like Facebook et al.
In the last few decades, tech companies have made billions off the data we've given them for free. Now, companies like Datacoup allow users to sell their data directly. Spark host Nora Young speaks with Datacoup founder Matt Hogan. She also speaks with Bart Custers, an Associate professor of law and digital technology at Leiden University in the Netherlands, who studies the value of our personal data to tech companies like Google and Facebook.
Taking tech apart
Canadian photographer Todd McLellan likes to disassemble tech objects in order to reveal their inner workings. He talks to Noar Young about his new book, Things Come Apart 2.0, which features 50 everyday items, from a Blackberry to an Amazon Echo, with each component neatly laid out with precision.
Attacking automated systems by 'tricking' AI
Say you're riding along in your autonomous car. You see a slightly defaced stop sign ahead but to your car's image recognition system, it looks like a speed limit sign, and the car sails right through the stop! Spark host Nora Young speaks with cybersecurity researcher Dawn Song, who has been able to demonstrate that AI systems can be fed malicious images or other inputs by attackers, with very scary results.