IBM developed NYPD surveillance tools that let cops pick targets based on skin color

The NYPD's secretive Lower Manhattan Security Coordination Center uses software from IBM in its video analytics system, which allows cops to automatically scan surveillance footage for machine-generated labels that identify clothing and other identifying classifiers. Read the rest

AI will try to paint what you tell it to, often generating surreal horrors

A research team wrote about how they trained a machine-learning AI to generate images from text descriptions. When fed birds as its dataset, it got very good at painting birds...

... But the more you feed it, the crazier it gets.

More collected here, and you can try it yourself thanks to Chris Valenzuela's online implementation.

Here are my efforts.

Read the rest

Here's the funniest, most scathing, most informative and most useful talk on AI and security

James Mickens (previously) has a well-deserved reputation for being the information security world's funniest speaker, and if that were all he did, he would still be worth listening to. Read the rest

A machine learning system trained on scholarly journals could correct Wikipedia's gendered under-representation problem

Quicksilver is a machine-learning tool from AI startup Primer: it used 30,000 Wikipedia entries to create a model that allowed it to identify the characteristics that make a scientist noteworthy enough for encyclopedic inclusion; then it mined the academic search-engine Semantic Scholar to identify the 200,000 scholars in a variety of fields; now it is systematically composing draft Wikipedia entries for scholars on its list who are missing from the encyclopedia. Read the rest

There's something eerie about bots that teach themselves to cheat

One of the holy grails of computer science is unsupervised machine learning, where you tell an algorithm what goal you want it to attain, and give it some data to practice on, and the algorithm uses statistics to invent surprising ways of solving your problem. Read the rest

The Russian equivalent to Alexa is a "good girl" but not too friendly, and is totally OK with wife-beating

Yandex is Russia's answer to Weibo, an everything-under-one-(semi-state-controlled)-roof online service, and its answer to Alexa is Alisa. Read the rest

Voice assistants suck, but they suck worse if you have an "accent"

Research into the shittiness of voice assistants zeroed in on a problem that many people were all-too-aware of: the inability of these devices to recognize "accented" speech ("accented" in quotes because there is no one formally correct English, and the most widely spoken English variants, such as Indian English, fall into this "accented" category). Read the rest

Several experts explain key ethical issues about AI

Artificial intelligence has nearly unimaginable potential to shape the world, but it poses a number of significant ethical questions that need to be carefully examined at every step to reduce bias. Several experts give a rundown of the main concerns. Read the rest

Voice assistants suck (empirically)

New research from legendary usability researchers The Nielsen (previously) Norman (previously) Group finds that voice assistants are basically a hot mess that people only use because they are marginally better than nothing. Read the rest

Google Translate's deep dream: some translation requests yield weird religious prophesies

Feed 19 repetitions of the word "dog" to Google Translate and ask it for a Maori conversion and you get this: "Doomsday Clock is three minutes at twelve. We are experiencing characters and a dramatic developments in the world, which indicate that we are increasingly approaching the end times and Jesus' return." Read the rest

Cool HOWTO on creating an AI to play 2048

The always entertaining and inspiring YouTuber Code Bullet decided to build a 2048-playing AI from scratch. It's really interesting to see him also build 2048 as part of the process. Read the rest

These nine companies are snapping up almost all the AI startups

In the current acquisition binge around artificial intelligence, tech behemoths with deep pockets lead the way, including Google, Apple, Facebook, Amazon, Intel, Microsoft, Twitter, and Salesforce. The only one with a limited consumer-facing presence is social monitoring firm Meltwater. Read the rest

Dank Learning: teaching a machine learning algorithm to generate memes

A physics student and an engineering student from Stanford fed 400,000 memes to a Long Short-Term Memory Recurrent Neural Network and asked it to generate more memes of its own. Read the rest

The Freeze-Frame Revolution: mutineers unstuck in time, strung out across an aeon

Peter Watts (previously) is a brilliant bastard of a science fiction writer, whose grim scenarios are matched by their scientific speculation; in his latest, a novella called The Freeze-Frame Revolution, Watts imagines a mutiny that stretches out across aeons, fought against a seemingly omnipotent AI.

The new Deep Fakes are in and they're spookily good

SIGGRAPH is coming, when all the amazeballs graphics research drops, and the previews are terrifying and astonishing by turns (sometimes both!). Read the rest

Machine learning may be most useful in tiny, embedded, offline processors

The tiny embedded processors in smart gadgets -- including much of the Internet of Shit -- are able to do a lot of sensing without exhausting their batteries, because sensing is cheap in terms of power consumption. Read the rest

Mozilla announces $225,000 in grants for "artwork and advocacy exploring AI"

Kevin from Mozilla writes, "In a world where biased algorithms, skewed data sets, and broken recommendation engines can radicalize YouTube users, promote racism, and spread fake news, it’s more important than ever to support artwork and advocacy work that educates and engages internet users." Read the rest

More posts