Roundup Another week has passed in the AI world, with military shenanigans, US government regulation and new chips from Qualcomm. Here’s the best of the rest. Are facial recognition coming to police body cams soon?:Axon, the largest supplier of body cameras to US police, has filed patents to develop, surprise, surprise, facial recognition software. It
Are facial recognition coming to police body cams soon?:Axon, the largest supplier of body cameras to US police, has filed patents to develop, surprise, surprise, facial recognition software.
It looks like CEO Rick Smith has changed his tune. Last year, during its Q2 earnings call Smith criticised the accuracy of facial recognition tech and said Axon weren’t actively pursuing it.
“This is technology that we don’t believe the accuracy thresholds are right where they need to be to make operational decisions off of facial recognition,” he said.
Fast forward seven months, and it’s all changed. Now, Axon has submitted three different patents for facial recognition, according to the FT. One patent is for a system to automatically detect and blur out personal information, probably for things like license plates and such, another matches faces to ones in a database, and the third detects objects like guns in video feeds. The software also analyses faces to guess the person’s age, race and gender.
A spokesperson said: “Filing a patent application is by no means a declaration of taking action or considering technology or product development. Our redaction product is currently limited to detection of faces. We do not match the detected faces against a centralised database which would be identification of the individual’s faces.”
How to think about AI and Jesus: The Ethics & Religious Liberty Commission, the public policy arm of America’s largest Christian denomination, the Southern Baptist Convention, is here to help if you have any existential dread from the emergence of AI.
AI is often painted as some machine that will become eventually become super intelligent enough that it’ll have profound impacts on mankind forever. Some of the top research labs like OpenAI and DeepMind are not shy about their ambitions to develop ‘artificial general intelligence”. So how do you reconcile the existence of all-knowing being with, erm, God?
The guidelines published by the Ethics & Religious Liberty Commission (ERLC) might seem pretty silly at first. But, perhaps surprisingly, it does discuss some of the most interesting technical aspects of AI right now, including the creation of sex dolls, AI in healthcare, bias and data and privacy.
Here are some of the juiciest parts. First of all, humanising AI bots is a no no. Technology shouldn’t be “assigned a level of human identity, worth, dignity, or moral agency”. The design of AI isn’t morally neutral and it should be used in a way to alleviate human suffering. Y’know love thy neighbor and all that.
Good news Christians. AI in medicine is acceptable if it enhances healthcare, but using it as a means to improve or change the actual body isn’t. So, no bionic humans or anything like that.
What about sexbots? Unfortunately, the ERLC doesn’t believe in any other kind of sex besides that between a husband and wife. It also doesn’t want AI to take all jobs as humanity shouldn’t “move towards life of pure leisure”.
The ERLC also disapproves of using AI to spurn out fake news or doctor images and videos like ‘deepfakes’. Finally, it warns against the misuse of data. “Data should not be mishandled, misused, or abused for sinful purposes to reinforce bias, strengthen the powerful, or demean the weak.” Take note Facebook and Google.
“In light of existential questions posed anew by the emergent technology of artificial intelligence (AI), we affirm that God has given us wisdom to approach these issues in light of Scripture and the gospel message.
“Christians must not fear the future or any technological development because we know that God is, above all, sovereign over history, and that nothing will ever supplant the image of God in which human beings are created. We recognize that AI will allow us to achieve unprecedented possibilities, while acknowledging the potential risks posed by AI if used without wisdom and care,” it said.
You can read the whole statement in full, here.
GAN research overview: Generative adversarial networks have exploded in popularity and it’s difficult to keep up with all the research.
There are constantly new papers that boast increasingly realistic looking images or ways to produce adversarial examples to trick other neural networks into misclassifying images, text, or audio. To help you pick out interesting or novel research among the noise, a computer scientist working at Google Brain has written about some of the important questions that are still unknown about GANs.
For example, how can GANs scale up for things other than producing flashy images? Why is it more difficult for text? Where should GANs be applied? How do GANs relate to adversarial examples?
It’s also a good overview on existing GAN research, if you’re interested give the Distill publication a read.
Basketball robot: Toyota showcased its basketball-playing robot Cue during a match in Japan.
You can watch a clip of it shooting a basketball from a pretty impressive distance for fun.
— ESPN (@espn) April 11, 2019
But like most robots, it’s pretty immobile so it probably can’t run, dribble, or tackle yet. ®
Becoming a Pragmatic Security Leader