Daily Archives: November 4, 2020

Police Will Pilot a Program to Live-Stream Amazon Ring Cameras

Posted on by

This is not a drill. Red alert: The police surveillance center in Jackson, Mississippi, will be conducting a 45-day pilot program to live stream the Amazon Ring cameras of participating residents.

Since Ring first made a splash in the private security camera market, we’ve been warning of its potential to undermine the civil liberties of its users and their communities. We’ve been especially concerned with Ring’s 1,000+ partnerships with local police departments, which facilitate bulk footage requests directly from users without oversight or having to acquire a warrant.

While people buy Ring cameras and put them on their front door to keep their packages safe, police use them to build comprehensive CCTV camera networks blanketing whole neighborhoods. This  serves two police purposes. First, it allows police departments to avoid the cost of buying surveillance equipment and to put that burden onto consumers by convincing them they need cameras to keep their property safe. Second, it evades the natural reaction of fear and distrust that many people would have if they learned police were putting up dozens of cameras on their block, one for every house.

Now, our worst fears have been confirmed. Police in Jackson, Mississippi, have started a pilot program that would allow Ring owners to patch the camera streams from their front doors directly to a police Real Time Crime Center. The footage from your front door includes you coming and going from your house, your neighbors taking out the trash, and the dog walkers and delivery people who do their jobs in your street. In Jackson, this footage can now be live streamed directly onto a dozen monitors scrutinized by police around the clock. Even if you refuse to allow your footage to be used that way, your neighbor’s camera pointed at your house may still be transmitting directly to the police.

[…]

Source: Police Will Pilot a Program to Live-Stream Amazon Ring Cameras | Electronic Frontier Foundation

Carbon footprint for ‘training GPT-3’ AI same as driving to the moon and back

Posted on by

Training OpenAI’s giant GPT-3 text-generating model is akin to driving a car to the Moon and back, computer scientists reckon.

More specifically, they estimated teaching the neural super-network in a Microsoft data center using Nvidia GPUs required roughly 190,000 kWh, which using the average carbon intensity of America would have produced 85,000 kg of CO2 equivalents, the same amount produced by a new car in Europe driving 700,000 km, or 435,000 miles, which is about twice the distance between Earth and the Moon, some 480,000 miles. Phew.

This assumes the data-center used to train GPT-3 was fully reliant on fossil fuels, which may not be true. The point, from what we can tell, is not that GPT-3 and its Azure cloud in particular have this exact scale of carbon footprint, it’s to draw attention to the large amount of energy required to train state-of-the-art neural networks.

The eggheads who produced this guesstimate are based at the University of Copenhagen in Denmark, and are also behind an open-source tool called Carbontracker, which aims to predict the carbon footprint of AI algorithms. Lasse Wolff Anthony, one of Carbontracker’s creators and co-author of a study of the subject of AI power usage, believes this drain on resources is something the community should start thinking about now, as the energy costs of AI have risen 300,000-fold between 2012 and 2018, it is claimed.

[…]

Source: AI me to the Moon… Carbon footprint for ‘training GPT-3’ same as driving to our natural satellite and back • The Register