skynettoday / skynet-today Goto Github PK

A deep look that somehow covers aspects other things miss

Get Table of contents to work. In theory easy w Jekyll, but not functional atm. Want it for about page esp.

http://www.mobihealthnews.com/content/study-ai-and-smartphone-based-device-help-identifying-diabetic-retinopathy

https://www.extremetech.com/extreme/261491-google-neural-network-can-predict-health-status-retina

http://www.pewglobal.org/2018/09/13/in-advanced-and-emerging-economies-alike-worries-about-job-automation/

jerry kaplan

https://www.youtube.com/watch?v=7Pq-S557XQU&t=531s
https://www.youtube.com/watch?v=hoDxcO2EOHM

https://news.ycombinator.com/item?id=16813823

Rebuke the whole 'horses' were not useful any more story - https://youtu.be/7Pq-S557XQU

Google released two audio recording of a conversation between an AI-powered assistant and human service providers. The result was amazing: from deceleration about Turing test passing to meditations about what it means about human nature. This is a classic case of hype in a teacup, but there are a few take-aways: 1) hold your horses, we're not there yet. They didn't even give a live-demo 2) google could have been more clear to begin with, though to be honest what we have heard IS amazing. However, companies should be responsible about not over-hyping either 3) if people contemplate the essence of our humanity after two calls to make reservations... what's in store when AIs do become more human-like? We need to start this discussion, soberly, about now, and brace ourselves for the future (while being quite excited about it too!)

Just create MD file w attribution and appropriate formatting

As covered in the digest:

Amazon’s Face Recognition Falsely Matched 28 Members of Congress With Mugshots

Jacob Snow, ACLU
ACLU matches image
Credit: from ACLU's article

The ACLU (American Civil Liberties Union) recently conducted a test of Amazon’s “Rekognition” facial recognition tool, and found that “the software incorrectly matched 28 members of Congress, identifying them as other people who have been arrested for a crime.” Their piece about the test goes on to say:

Matching people against arrest photos is not a hypothetical exercise. Amazon is aggressively marketing its face surveillance technology to police, boasting that its service can identify up to 100 faces in a single image, track people in real time through surveillance cameras, and scan footage from body cameras. A sheriff’s department in Oregon has already started using Amazon Rekognition to compare people’s faces against a mugshot database, without any public debate.

And that:

Congress must take these threats seriously, hit the brakes, and enact a moratorium on law enforcement use of face recognition. This technology shouldn’t be used until the harms are fully considered and all necessary steps are taken to prevent them from harming vulnerable communities.

Amazon’s Dr. Matt Wood wrote a response that noted that “The ACLU has not published its data set, methodology, or results in detail, so we can only go on what they’ve publicly said” and listed several issues with the bits of the methodology that were disclosed (such as the confidence threshold, training dataset, and more). The ACLU responded by denying were there issues with their methodology and stating that “In its five stages of grief over its dangerous face surveillance product, Amazon is clearly stuck at denial”.

The motherboard NMT story

Discuss how good a concept exponentialism is https://www.technologyreview.com/s/609048/the-seven-deadly-sins-of-ai-predictions/

Write news post, maybe start thinking about editorial possibilities

Just do a pass and see if anything looks off , then I'll take the updated draft and finalize

A non-snarky post about the very real and troubling ability to overlay anyone's face on porn.

Starting point: https://motherboard.vice.com/amp/en_us/article/neqb98/reddit-shuts-down-deepfakes?__twitter_impression=true

Some thoughts:

• Could have a sentence highlighting how this is the sort of stuff people should legitimately be worried about wrt AI, and not silly stuff (in fact, should probably hint at that in headline)
• Could actually spend some time discussing how media coverage mostly got it right
• Maybe can include a gif/image for demonstration, though definitely of a non-racy nature

Add support for longer and more in depth coverage not necessarily around one event. Will be a toggle sort of thing between news and editorials.

https://www.theatlantic.com/magazine/archive/2018/06/henry-kissinger-ai-could-mean-the-end-of-human-history/559124/
https://twitter.com/etzioni/status/996623076711264256
https://www.facebook.com/yann.lecun/posts/10155284356512143

┆Attachments: Draft 2: How close are we to solving translation with neural networks

First one, so will have to work a bit extra on this one

And posts should be renamed to Coverage

1. Ideally want something similar to distill.pub
2. in the briefs section we could have a logo/icon beside each article to indicate whether it represents good quality journalism or AI hype. Ex: rotten tomato for articles that fall in the AI hype category.

Think Answer or the singularity vs. model bias, privacy concerns, adversarial examples, and misuses (deepfakes etc.).

We've been seeing endless headlines containing the word AI while actually referring to relatively trivial ML or other techniques. Many of them seem to imply that AI is a magical entity that just keeps on amassing capabilities. That is misleading, and a big factor in creating and pushing forward the hype that we see as harmful and dangerous. It is time we be more careful with how we define technologies and refer to them publicly.

Got to home screen
Click 'Editorials'
Click on latest editorial
Click back buttion

Briefs will be displayed despite editorials section being highlighted

1. Talk to experts
2. Don't say an AI (link to Limor's longer piece)
3. No clickbait titles

┆Attachments: skynettoday/skynet-today: Issue #78

Kind of an overview, maybe some about hype/what problems it can solve

https://github.com/nathancy/jekyll-embed-video

http://listjs.com/examples/pagination/

Really silly stort, basically unethical portrayal if robot by researchers
Quote Yann lecunn and media fact checks