The Sunday Artificial Intelligence News Review: 18th June 2017
by Michele Baker
The temperatures are soaring here in the beautiful Sussex countryside, and I am spending the weekend hanging out in my mum’s garden under the apple trees while Indie rides her bike and Billy explores the surrounding fields and woodland. It could not be more idyllic; I plan to get back to it as soon as possible. So with no further ado, I’m going to crack on with the business of this week’s artificial intelligence news.
To start off, I am pleased to say that my recent TDMB Presents… podcast, featuring Simon Gosling from Unruly talking about the company’s concept smart home and the next era of ambient advertising in the connected future, is now live. You can listen to it here:
You can also catch a full rundown of mine and Jon’s experience of the Unruly concept home and offices in my summarising blog post, complete with video inside the smart home itself.
Right, so – onto the biggest artificial intelligence news this week:
I seem to remember that Google’s AI did this a while back, but now Facebook reports that its chatbots converse with one another in a language of their own. The researchers wrote that they had to tweak one of their models to prevent the bot-to-bot conversation diverging “from human language as the agents developed their own language for negotiating”. It might all sound a bit terrifying, but Silicon Valley and the media are quick to reassure us that it’s cool. Nothing to see here, move along…
Whether AI developing their own language that we cannot understand is a sign of the impending robopocalypse or not, it does redefine our ideas of what is exclusively human – and reminds us of just how little we really know.
Oh look, it’s everyone’s favourite vomit-inducing, right-wing trashpile – Breitbart! This article chews over the news that Facebook is trying to counter extremist content on its website using artificial intelligence, which should be considered, y’know, a good thing, right? But, in predictable Breitbart fashion, the publication is quick to condemn Facebook for removing pages like Menimist (massive eye-roll) and other men’s rights activist pages, as well as anti-ISIS and anti-Islamist pages. They also draw attention to the fact that the Black Panther Party of Mississippi was allowed to keep a cartoon of a police officer having his throat slit. Of course, they fail to mention the countless disgusting pages of racist, homo- and transphobic, and misogynistic pages, and content showing sickening sexual and animal abuse that haven’t been taken down. Because that wouldn’t fit their narrative. Anyway, let’s move on.
In an effort to lower the barrier for entry to new companies wanting to develop commercial AI solutions, the Montreal-based platform and incubator, Element AI, has announced a round of Series A funding amounting to a whopping $102m.
PICK OF THE WEEK
Nice video here of WikiLeaks editor, Julian Assange, speaking about the threat of AI-controlled social media at a panel discussion that included Slovenian philosopher (and so-called Elvis of Cultural Theory, Slavoj Žižek. “Imagine a Daily Mail run by essentially Artificial Intelligence, what does that look like when there’s only the Daily Mail worldwide?” Assange said. “That’s what Facebook and Twitter will shift into.” It’s a really interesting point, and one we need to talk about: will AI become some horrible propaganda machine? It really could do, and mainstream media is bad enough without adding AI to the pot.
Google DeepMind researchers have developed a simple algorithm that can reason at a human level (whatever that may be!) and has already beaten humans at a complex image comprehension test. The results are really quite impressive… I really recommend giving this one a thorough read.
I know I usually give you my favourite artificial intelligence news articles from my inbox here every week, but ‘I’m sorry, Dave, I’m afraid I can’t do that’ – for two reasons. The first is that there was literally nothing worth sharing with you. I am not going to just go ahead and link you up with any old crap… that’s not how I work. Unless you want to read about why Demi Moore’s front teeth fell out, the ten biggest theories behind Tupac’s death, or the science of making friends, none of which have anything to do with AI, then I have nothing for you today. Sorry about that. The second reason is that I am not going to trawl the web for titbits when there’s a whole beautiful day outside and my loved-ones waiting for me. Yours are probably waiting for you, too. So go to them, go get that Vitamin D, go live your life.
I’ll be back.