Author: Natasha Lomas

Prisma’s style transfer tech creeps into kids’ books

TECHCRUNCH The style transfer craze kicked off by an app called Prisma a couple of years ago led to a tsunami of painterly selfies flooding social feeds for several months, as we reported at the time, before the rapacious, face-snapping hoards shifted their attention toward fresh spectacles. But that’s not the end of the story. The same tech is now creeping into (paper) kids’ books, via a partnership between children’s publisher startup, Kabook, and Prisma Labs: aka the b2b entity that the original app makers pivoted to in late 2017. So instead of AI sending robots into a human-slaying frenzy, per the usual dystopian sci-fi storyline, we find ourselves confronted with neural nets being used to serve up contextual illustrations of children so parents can gift personalized books that seamlessly insert a child’s likeness into the story, thereby casting them as a character in the tale. Not the end of the world then. Well, not unless you view this kind of self-centered content manipulation as a threat to children’s imaginations and developing sense of empathy. (The research on any ‘little princes in training’ will, unfortunately, have to wait a few decades to come through though.) The Kabook integration is the first consumer product partnership that Prisma Labs has scored, according to a press release from the pair. And while they note there are other publishing services that offer the chance to...

Read More

Facebook finds and kills another 512 Kremlin-linked fake accounts

TECHCRUNCH Two years on from the U.S. presidential election, Facebook continues to have a major problem with Russian disinformation being megaphoned via its social tools. In a blog post today the company reveals another tranche of Kremlin-linked fake activity — saying it’s removed a total of 471 Facebook pages and accounts, as well as 41 Instagram accounts, which were being used to spread propaganda in regions where Putin’s regime has sharp geopolitical interests. In its latest reveal of “coordinated inauthentic behavior” — aka the euphemism Facebook uses for disinformation campaigns that rely on its tools to generate a veneer of authenticity and plausibility in order to pump out masses of sharable political propaganda — the company says it identified two operations, both originating in Russia, and both using similar tactics without any apparent direct links between the two networks. One operation was targeting Ukraine specifically, while the other was active in a number of countries in the Baltics, Central Asia, the Caucasus, and Central and Eastern Europe. “We’re taking down these Pages and accounts based on their behavior, not the content they post,” writes Facebook’s Nathaniel Gleicher, head of cybersecurity policy. “In these cases, the people behind this activity coordinated with one another and used fake accounts to misrepresent themselves, and that was the basis for our action.” Sputnik link Discussing the Russian disinformation op targeting multiple countries, Gleicher says Facebook found what looked...

Read More

Facebook urged to give users greater control over what they see

TECHCRUNCH Academics at the universities of Oxford and Stanford think Facebook should give users greater transparency and control over the content they see on its platform. They also believe the social networking giant should radically reform its governance structures and processes to throw more light on content decisions, including by looping in more external experts to steer policy. Such changes are needed to address widespread concerns about Facebook’s impact on democracy and on free speech, they argue in a report published today which includes a series of recommendations for reforming Facebook (entitled: Glasnost! Nine Ways Facebook Can Make Itself a Better Forum for...

Read More

Most Facebook users still in the dark about its creepy ad practices, Pew finds

TECHCRUNCH A study by the Pew Research Center suggests most Facebook users are still in the dark about how the company tracks and profiles them for ad-targeting purposes. Pew found three-quarters (74%) of Facebook users did not know the social networking behemoth maintains a list of their interests and traits to target them with ads, only discovering this when researchers directed them to view their Facebook ad preferences page. A majority (51%) of Facebook users also told Pew they were uncomfortable with Facebook compiling the information. While more than a quarter (27%) said the ad preference listing Facebook had...

Read More

Facebook is launching political ad checks in Nigeria, Ukraine, EU and India in coming months

TECHCRUNCH Facebook is launching some of its self-styled ‘election security’ initiatives into more markets in the coming months ahead of several major votes in countries around the world. In an interview with Reuters the social networking giant confirmed it’s launching checks on political adverts on its platform in Nigeria, Ukraine and the European Union, reiterating too that ad transparency measures will launch in India ahead of its general election. Although it still hasn’t confirmed how it will respond in other countries with looming votes this year, including Australia, Indonesia, Israel and the Philippines. Concern about election interference in the era...

Read More

Wrest control from a snooping smart speaker with this teachable “parasite”

TECHCRUNCH What do you get when you put one Internet connected device on top of another? A little more control than you otherwise would in the case of Alias the “teachable ‘parasite’” — an IoT project smart speaker topper made by two designers, Bjørn Karmann and Tore Knudsen. The Raspberry Pi-powered, fungus-inspired blob’s mission is to whisper sweet nonsense into Alexa’s (or Google Home’s) always-on ear so it can’t accidentally snoop on your home. [embedded content] Project Alias from Bjørn Karmann on Vimeo. Alias will only stop feeding noise into its host’s speakers when it hears its own wake command — which can be whatever you like. The middleman IoT device has its own local neural network, allowing its owner to christen it with a name (or sound) of their choosing via a training interface in a companion app. The open source TensorFlow library was used for building the name training component. So instead of having to say “Alexa” or “Ok Google” to talk to a commercial smart speaker — and thus being stuck parroting a big tech brand name in your own home, not to mention being saddled with a device that’s always vulnerable to vocal pranks (and worse: accidental wiretapping) — you get to control what the wake word is, thereby taking back a modicum of control over a natively privacy-hostile technology. This means you could rename Alexa...

Read More

German court tosses Qualcomm’s latest iPhone patent suit

TECHCRUNCH Qualcomm has had a patent lawsuit against Apple dismissed by a court in Mannheim, Germany, as groundless (via Reuters). The chipmaker had argued Intel-powered iPhones infringed a transistor switch patent it holds. But in an initial verbal decision the court disagreed. Qualcomm has said it will appeal. In a statement, Don Rosenberg, Qualcomm’s executive VP and general counsel, said: “Apple has a history of infringing our patents. While we disagree with the Mannheim court’s decision and will appeal, we will continue to enforce our [intellectual property] rights against Apple worldwide.” We’ve reached out to Apple for comment. The...

Read More

Testing times for second wave scooter startups

TECHCRUNCH Investors are still pouring millions into scooter startups, albeit sometimes at flat valuations. At the same time a little cash is flowing the other way, in cases where cities have realized the importance of prioritizing the needs of the local environment and its citizens, over and above the ambitions of VCs for a swift and lucrative exit. Scooter startups affected by such regulatory bumps in the road are, unsurprisingly, rather less keen to shout about this sort of policy friction and the negative cash and ride flow it generates. In one recent incident in Spain, in the Catalan capital of Barcelona, El Pais reported that the town hall fined a local scooter startup, called Reby, for contravening urban mobility rules. The startup is so new it doesn’t even have scooters available for public hire yet. But it’s already had some of its ‘test’ rides removed by police and been fined for breaking scooter sharing rules. If it was hoping to copy-paste from an Uber 1.0 playbook, things aren’t looking good for Reby. (Indeed, that’s a very tatty manual in most places these days.) Spain’s capital city Madrid also forced a temporary suspension on scooter sharing startups recently, as we reported last month, after changes to mobility laws that tighten the screw on scooter sharing — requiring already operational startups to tweak how their rides operate in order to come into compliance. While Madrid...

Read More

Right Now in Politics and Business