Enjoy this weekly feed of inspiring, interesting and intellectual articles and news, with a focus on disruption!
Deepfakes – The latest national security threat
Someone has trained an AI to basically do what was done with Brad Pitt in Benjamin Button, replace the face or head of a person in a video. The process has been given a name, ‘deepfake’, and just a few weeks after the release, it was recently considered as the latest considerable national security threat according to lawmakers. So far, this AI has been used (and I guess continuously trained) primarily for fake celebrity porn. Quite harmful but just imagine the potential applications for anyone with access to surveillance video (ANY state or country).
Another, related, AI project that is quite scary is one from OpenAI (Elon Musk co-founder) which has produced an AI that can create very convincing text samples. Sort of like internet (fake?) news, since it was trained by 8 million web pages.
“OpenAI is worried that bad actors would use automated text generators to whip up deceptive, biased, or abusive language at scale”
Combine these two (video and text) and you have the worst nightmare possible in terms of fake news. “What’s your source? Well, press play and see for yourself…”
Waymo leads the autonomous vehicle race
Data released from the California Department of Motor Vehicles shows that Waymo is well ahead of its competitors in terms of self-driving technology seen from how often a person has to intervene and control the car. Waymo’s test driver needs to intervene on average 0.09 times every 1000 miles, compared to the closest competitor GM (the only one close) whose figure is 0.19. Waymo can be about to become a key player in the future automotive industry.
Yet to be answered though, will LIDAR (Waymo and all others) beat cameras+AI (Tesla) to full autonomy?