Daily Archives: July 9, 2018

Controversial copyright law rejected by EU parliament

Posted on by

A controversial overhaul of the EU’s copyright law that sparked a fierce debate between internet giants and content creators has been rejected.

The proposed rules would have put more responsibility on websites to check for copyright infringements, and forced platforms to pay for linking to news.

A slew of high-profile music stars had backed the change, arguing that websites had exploited their content.

But opponents said the rules would stifle internet freedom and creativity.

The move was intended to bring the EU’s copyright laws in line with the digital age, but led to protests from websites and much debate before it was rejected by a margin of 318-278 in the European Parliament on Thursday.

What were they voting for?

The proposed legislation – known as the Copyright Directive – was an attempt by the EU to modernise its copyright laws, but it contained two highly-contested parts.

The first of these, Article 11, was intended to protect newspapers and other outlets from internet giants like Google and Facebook using their material without payment.

But it was branded a “link tax” by opponents who feared it could lead to problems with sentence fragments being used to link to other news outlets (like this).

Article 13 was the other controversial part. It put a greater responsibility on websites to enforce copyright laws, and would have meant that any online platform that allowed users to post text, images, sounds or code would need a way to assess and filter content.

The most common way to do this is by using an automated copyright system, but they are expensive. The one YouTube uses cost $60m (£53m), so critics were worried that similar filters would need to be introduced to every website if Article 13 became law.

There were also concerns that these copyright filters could effectively ban things like memes and remixes which use some copyrighted material.

Source: Controversial copyright law rejected by EU parliament – BBC News

Very glad to see common sense prevailing here. Have you ever thought about how strange it would  be if you could bill someone every time they read your email or your reports? How do musicians think it’s ok to bill people when they are not playing?

Former NSO Group Employee Accused of Stealing Phone Spy Tools

Posted on by

Israeli hacking firm NSO Group is mostly known for peddling top-shelf malware capable of remotely cracking into iPhones. But according to Israeli authorities, the company’s invasive mobile spy tools could have wound up in the hands of someone equally, if not far more, devious than its typical government clients.

A 38-year-old former NSO employee has been accused of stealing the firm’s malware and attempting to sell it for $50 million in cryptocurrency on the dark net, according to a widely reported indictment first published by Israeli press.

The stolen software is said to be worth hundreds of millions of dollars.

According to Israel’s Justice Ministry, the ex-employee was turned in by a potential buyer. The suspect was arrested on June 5, Reuters reported. The accused has been charged with employee theft, attempting to sell security tools without a license, and conduct that could harm state security

Source: Former NSO Group Employee Accused of Stealing Phone Spy Tools

Obviously security holes found will be exploited, which is why responsible disclosure is a good idea. It’s much better for devices to be secure than for intelligence agencies to be able to exploit holes – because non-nation state actors (read: criminals, although there are nations who think other nations are criminal) also have access to these holes.

App Traps: How Cheap Smartphones Siphon User Data in Developing Countries

Posted on by

For millions of people buying inexpensive smartphones in developing countries where privacy protections are usually low, the convenience of on-the-go internet access could come with a hidden cost: preloaded apps that harvest users’ data without their knowledge.

One such app, included on thousands of Chinese-made Singtech P10 smartphones sold in Myanmar and Cambodia, sends the owner’s location and unique-device details to a mobile-advertising firm in Taiwan called General Mobile Corp., or GMobi. The app also has appeared on smartphones sold in Brazil and those made by manufacturers based in China and India, security researchers said.

Taipei-based GMobi, with a subsidiary in Shanghai, said it uses the data to show targeted ads on the devices. It also sometimes shares the data with device makers to help them learn more about their customers.

Smartphones have been billed as a transformative technology in developing markets, bringing low-cost internet access to hundreds of millions of people. But this growing population of novice consumers, most of them living in countries with lax or nonexistent privacy protections, is also a juicy target for data harvesters, according to security researchers.

Smartphone makers that allow GMobi to install its app on phones they sell are able to use the app to send software updates for their devices known as “firmware” at no cost to them, said GMobi Chief Executive Paul Wu. That benefit is an important consideration for device makers pushing low-cost phones across emerging markets.

“If end users want a free internet service, he or she needs to suffer a little for better targeting ads,” said a GMobi spokeswoman.

[…]

Upstream Systems, a London-based mobile commerce and security firm that identified the GMobi app’s activity and shared it with the Journal, said it bought four new devices that, once activated, began sending data to GMobi via its firmware-updating app. This included 15-digit International Mobile Equipment Identification, or IMEI, numbers, along with unique codes called MAC addresses that are assigned to each piece of hardware that connects to the web. The app also sends some location data to GMobi’s servers located in Singapore, Upstream said.

Source: App Traps: How Cheap Smartphones Siphon User Data in Developing Countries – WSJ

 

I like the way even GMobi thinks users getting targetted advertising are suffering!

An AI system for editing music in videos can isolate single instruments

Posted on by

Amateur and professional musicians alike may spend hours pouring over YouTube clips to figure out exactly how to play certain parts of their favorite songs. But what if there were a way to play a video and isolate the only instrument you wanted to hear?

That’s the outcome of a new AI project out of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL): a deep-learning system that can look at a video of a musical performance, and isolate the sounds of specific instruments and make them louder or softer.

The system, which is “self-supervised,” doesn’t require any human annotations on what the instruments are or what they sound like.

Trained on over 60 hours of videos, the “PixelPlayer” system can view a never-before-seen musical performance, identify specific instruments at pixel level, and extract the sounds that are associated with those instruments.

For example, it can take a video of a tuba and a trumpet playing the “Super Mario Brothers” theme song, and separate out the soundwaves associated with each instrument.

The researchers say that the ability to change the volume of individual instruments means that in the future, systems like this could potentially help engineers improve the audio quality of old concert footage. You could even imagine producers taking specific instrument parts and previewing what they would sound like with other instruments (i.e. an electric guitar swapped in for an acoustic one).

Source: An AI system for editing music in videos | MIT News

DeepMind’s AI agents exceed ‘human-level’ gameplay in Quake III

Posted on by

AI agents continue to rack up wins in the video game world. Last week, OpenAI’s bots were playing Dota 2; this week, it’s Quake III, with a team of researchers from Google’s DeepMind subsidiary successfully training agents that can beat humans at a game of capture the flag.

As we’ve seen with previous examples of AI playing video games, the challenge here is training an agent that can navigate a complex 3D environment with imperfect information. DeepMind’s researchers used a method of AI training that’s also becoming standard: reinforcement learning, which is basically training by trial and error at a huge scale.

Agents are given no instructions on how to play the game, but simply compete against themselves until they work out the strategies needed to win. Usually this means one version of the AI agent playing against an identical clone. DeepMind gave extra depth to this formula by training a whole cohort of 30 agents to introduce a “diversity” of play styles. How many games does it take to train an AI this way? Nearly half a million, each lasting five minutes.

As ever, it’s impressive how such a conceptually simple technique can generate complex behavior on behalf of the bots. DeepMind’s agents not only learned the basic rules of capture the flag (grab your opponents’ flag from their base and return it to your own before they do the same to you), but strategies like guarding your own flag, camping at your opponent’s base, and following teammates around so you can gang up on the enemy.

To make the challenge harder for the agents, each game was played on a completely new, procedurally generated map. This ensured the bots weren’t learning strategies that only worked on a single map.

Unlike OpenAI’s Dota 2 bots, DeepMind’s agents also didn’t have access to raw numerical data about the game — feeds of numbers that represents information like the distance between opponents and health bars. Instead, they learned to play just by looking at the visual input from the screen, the same as a human. However, this does not necessarily mean that DeepMind’s bots faced a greater challenge; Dota 2 is overall a much more complex game than the stripped-down version of Quake III that was used in this research.

To test the AI agents’ abilities, DeepMind held a tournament, with two-player teams of only bots, only humans, and a mixture of bots and humans squaring off against one another. The bot-only teams were most successful, with a 74 percent win probability. This compared to 43 precent probability for average human players, and 52 percent probability for strong human players. So: clearly the AI agents are the better players.

A graph showing the Elo (skill) rating of various players. The “FTW” agents are DeepMind’s, which played against themselves in a team of 30.
Credit: DeepMind

However, it’s worth noting that the greater the number of DeepMind bots on a team, the worse they did. A team of four DeepMind bots had a win probability of 65 percent, suggesting that while the researchers’ AI agents did learn some elements of cooperative play, these don’t necessarily scale up to more complex team dynamics.

As ever with research like this, the aim is not to actually beat humans at video games, but to find new ways of teaching agents to navigate complex environments while pursuing a shared goal. In other words, it’s about teaching collective intelligence — something that has (despite abundant evidence to the contrary) been integral to humanity’s success as a species. Capture the flag is just a proxy for bigger games to come.

Source: DeepMind’s AI agents exceed ‘human-level’ gameplay in Quake III – The Verge

Mitsubishi Wants Your Driving Data, and It’s Willing to Throw in a Free Cup of Coffee to Get It

Posted on by

Automakers want in on the highly lucrative big data game and Mitsubishi is willing to pay for the privilege. In exchange for running the risk of jacking up its customers’ insurance premiums, the car manufacturer is offering drivers $10 off of an oil change and other rewards. Consumers will have to decide if a gift card is worth giving up their privacy.

According to the Wall Street Journal, Mitsubishi’s new smartphone app is the first of its kind. A driver can sign up and allow their driving habits to be tracked by their phone’s sensors, which monitor data points like acceleration, location, and rotation. Along the way, they’ll earn badges (reward points) based on good driving practices like staying under the speed limit. For now, the badges can be exchanged for discounted oil changes or car accessories, but the company plans to expand its incentives to other small perks like free cups of coffee by the end of the year.

It may seem like a win-win situation: You pay a little more attention to being a good driver and you get a little bonus for your efforts. But the first customer for all that data is State Auto Insurance Companies, which will be using it to create better risk models and adjust users’ premiums accordingly. It doesn’t appear that the data will be anonymized because the Journal reports that, after a trial period, insurers will be able to build a customer risk profile on users of the app that will then be used to determine rates. We reached out to Mitsubishi to ask about its anonymization of data but didn’t receive an immediate reply.

Mike LaRocco, State Auto’s CEO, framed this as a benefit to consumers when speaking with the Journal. “They’ll get a much more accurate quote from day one,” he claimed. That might be true, but it does nothing to assuage fears that insurance companies could penalize drivers who don’t voluntarily give up their data.

Ford also has an app that shares data with insurance companies, but it’s not offering any of those sweet, sweet gift cards. And at a moment when many people are debating whether tech giants should be paying us for our data, one could argue that Mitsubishi is doing the right thing. But as car companies are building web connectivity into their new models, we could easily see this become standard practice without offering drivers a choice or a reward. A study by McKinsey & Co from 2016, estimated that monetizing car data could be worth between $450-750 billion by 2030. Of course, autonomous vehicles could become more prevalent by then. And as long as they work as promised, insurance companies will be less necessary.

[Wall Street Journal]

Source: Mitsubishi Wants Your Driving Data, and It’s Willing to Throw in a Free Cup of Coffee to Get It

EU asks you to tell them if you want Daylight Savings Time

Posted on by

Objective of the consultation

Following a number of requests from citizens, from the European Parliament, and from certain EU Member States, the Commission has decided to investigate the functioning of the current EU summertime arrangements and to assess whether or not they should be changed.

In this context, the Commission is interested in gathering the views of European citizens, stakeholders and Member States on the current EU summertime arrangements and on any potential change to those arrangements.

How to submit your response

The online questionnaire is accessible in all official EU languages (except Irish) and replies may be submitted in any EU language. We do encourage you to answer as much as possible in English though.

You may pause at any time and continue later. Once you have submitted your answers, you can download a copy of your completed responses.

Source: Public Consultation on summertime arrangements | European Commission

Versius Robot allows keyhole surgery to be performed with 1/2 hour training instead of 80 sessions

Posted on by

It is the most exacting of surgical skills: tying a knot deep inside a patient’s abdomen, pivoting long graspers through keyhole incisions with no direct view of the thread.

Trainee surgeons typically require 60 to 80 hours of practice, but in a mock-up operating theatre outside Cambridge, a non-medic with just a few hours of experience is expertly wielding a hook-shaped needle – in this case stitching a square of pink sponge rather than an artery or appendix.

The feat is performed with the assistance of Versius, the world’s smallest surgical robot, which could be used in NHS operating theatres for the first time later this year if approved for clinical use. Versius is one of a handful of advanced surgical robots that are predicted to transform the way operations are performed by allowing tens or hundreds of thousands more surgeries each year to be carried out as keyhole procedures.

[…]

The Versius robot cuts down the time required to learn to tie a surgical knot from more than 100 training sessions, when using traditional manual tools, to just half an hour, according to Slack.

[…]

Versius comprises three robotic limbs – each slightly larger than a human arm, complete with shoulder, elbow and wrist joints – mounted on bar-stool sized mobile units.

Controlled by a surgeon at a console, the limbs rise, fall and swivel silently and smoothly. The robot is designed to carry out a wide range of keyhole procedures, including hysterectomies, prostate removal, ear, nose and throat surgery, and hernia repair. CMR claims the costs of using the robot will not be significantly higher than for a conventional keyhole procedure.

Source: The robots helping NHS surgeons perform better, faster – and for longer | Society | The Guardian

Fitness app Polar even better at revealing secrets than Strava and Garmin

Posted on by

Online investigations outfit Bellingcat has found that fitness tracking kit-maker Polar reveals both the identity and daily activity of its users – including soldiers and spies.

Many users of Polar’s devices and app appear not to have paid attention to their privacy settings, as a result a Bellingcat writer found 6,460 individuals from 69 countries. More than 200 of them left digital breadcrumbs around sensitive locations.

Bellingcat’s report claimed the Polar Flow social-fitness site produces more compromising data than other fitness-trackers than previous leaks: “Compared to the similar services of Garmin and Strava, Polar publicizes more data per user in a more accessible way, with potentially disastrous results.“

“Tracing all of this information is very simple through the site: find a military base, select an exercise published there to identify the attached profile, and see where else this person has exercised.”

Bellingcat notes that the big difference between Polar and Strava is that the former offers more comprehensive data, more easily, covering everything a user has uploaded to the platform since 2014.

Source: Fitness app Polar even better at revealing secrets than Strava • The Register

Open plan offices flop – you talk less, IM more, if forced to flee a cubicle

Posted on by

Open plan offices don’t deliver their promised benefits of more face-to-face collaboration and instead make us misanthropic recluses and more likely to use electronic communications tools.

So says a new article in the Philosophical Transactions of the Royal Society B, by Harvard academics Ethan S. Bernstein, Stephen Turban. The pair studied two Fortune 500 companies that adopted open office designs and wrote up the results as “The impact of the ‘open’ workspace on human collaboration”.

[…]

Analysis of the data revealed that “volume of face-to-face interaction decreased significantly (approx. 70%) in both cases, with an associated increase in electronic interaction.”

“In short, rather than prompting increasingly vibrant face-to-face collaboration, open architecture appeared to trigger a natural human response to socially withdraw from officemates and interact instead over email and IM.”

In the first workplace studied, “IM message activity increased by 67% (99 more messages) and words sent by IM increased by 75% (850 more words). Thus — to restate more precisely — in boundaryless space, electronic interaction replaced F2F interaction.”

The second workplace produced similar results.

The authors reach three conclusions, the first of which is that open offices “can dampen F2F interaction, as employees find other strategies to preserve their privacy; for example, by choosing a different channel through which to communicate.”

Source: Open plan offices flop – you talk less, IM more, if forced to flee a cubicle • The Register

Empathic AI (Dutch)

Posted on by

Bedrijven worden emotioneler: gebruikersinterfaces, chatbots en andere componenten zijn steeds beter in staat om de emotionele staat van gebruikers in te schatten en emotie te simuleren als ze terug praten. Volgens een Gartner-rapport eerder dit jaar weten apparaten over vier jaar “meer over je emotionele staat dan je eigen familie”.

Herkennen van emotie

Deep learning kan geavanceerd emoties herkennen zoals geluk, verrassing, woede, verdriet, angst en afschuw – tot meer dan twintig subtielere emoties zoals bewondering, blije verrassing en haat. (Psychologen beweren dat mensen 27 verschillende emoties hebben.)

De Universiteit van Ohio ontwikkelde een programma dat 21 emoties herkent op basis van gezichtsuitdrukkingen op foto’s. Het schokkende: De onderzoekers beweren dat hun systeem deze emoties beter detecteert dan mensen. Er is een goede reden en een geweldige reden voor emotionele interfaces in de organisatie.

Kwaliteitsinteracties

Ten eerste de goede reden. De “empathie economie” is de monetaire of zakelijke waarde die door AI wordt gecreëerd en die menselijke emoties detecteert en simuleert, een vermogen dat klantenservice, virtuele assistenten, robotica, fabrieksveiligheid, gezondheidszorg en transport zal transformeren.

Uit een Cogito-onderzoek van Frost & Sullivan gaf 93% van de ondervraagden aan dat interacties met de klantenservice van invloed zijn op hun perceptie van een bedrijf. En empathie is één van de belangrijkste factoren in kwaliteitsinteracties, volgens het bedrijf. Cogito’s AI-software, die uitgebreid is gebaseerd op gedragswetenschappelijk onderzoek van MIT’s Human Dynamics Lab, analyseert de emotionele toestand van klanten en geeft directe feedback aan menselijke call center agents, waardoor ze gemakkelijker meevoelen met klanten.

Zorg en andere toepassingen

Dit soort technologie geeft callcentermedewerkers empathische vermogens, die de publieke perceptie van een bedrijf sterk kunnen verbeteren. Bedrijven als Affectiva en Realeyes bieden cloud-gebaseerde oplossingen die webcams gebruiken om gezichtsuitdrukkingen en hartslag te volgen (door de polsslag in de huid van het gezicht te detecteren). Een van de toepassingen is marktonderzoek: consumenten kijken naar advertenties, en de technologie detecteert hoe ze denken over de beelden of woorden in de advertenties.

De ondernemingen zijn op zoek naar andere gebieden, zoals de gezondheidszorg, waar geautomatiseerde call centers depressie of pijn in de stem van de beller zou kunnen detecteren, zelfs als de beller niet in staat is deze emoties uit te drukken.

Stemming detecteren

Een robot met de naam Forpheus, gemaakt door Omron Automation in Japan en gedemonstreerd tijdens CES in januari, speelt pingpong. Een deel van haar arsenaal van tafeltennisvaardigheden is haar vermogen om lichaamstaal te lezen om zowel de stemming en vaardigheid niveau van de menselijke tegenstander te achterhalen.

Het gaat natuurlijk niet om pingpong, maar het doel is industriële machines die “in harmonie” met de mens werken, wat zowel de productiviteit als de veiligheid verhoogt. Door bijvoorbeeld de lichaamstaal van fabrieksarbeiders te lezen, konden industriële robots anticiperen op hoe en waar mensen zich zouden kunnen bewegen.

Source: Empathische AI komt eraan – en dat is mooi – Computerworld

Nostalgic social network ‘Timehop’ loses data from 21 million users

Posted on by

A service named “Timehop” that claims it is “reinventing reminiscing” – in part by linking posts from other social networks – probably wishes it could go back in time and reinvent its own security, because it has just confessed to losing data describing 21 million members and can’t guarantee that the perps didn’t slurp private info from users’ social media accounts.

“On July 4, 2018, Timehop experienced a network intrusion that led to a breach of some of your data,” the company wrote. “We learned of the breach while it was still in progress, and were able to interrupt it, but data was taken.”

Names and email addresses were lifted, as were “Keys that let Timehop read and show you your social media posts (but not private messages)”. Timehop has “deactivated these keys so they can no longer be used by anyone – so you’ll have to re-authenticate to our App.”

The breach also led to the loss of access tokens Timehop uses to access other social networks such as Twitter, Facebook and Instagram and the posts you’ve made there. Timehop swears blind that the tokens have been revoked and just won’t work any more.

But the company has also warned that “there was a short time window during which it was theoretically possible for unauthorized users to access those posts” but has “no evidence that this actually happened.”

It can’t be as almost-comforting on the matter of purloined phone numbers, advising that for those who shared such data with the company “It is recommended that you take additional security precautions with your cellular provider to ensure that your number cannot be ported.” Oh thanks for that, Timehop. And thanks, also, for not using two-factor authentication, because that made the crack possible. “The breach occurred because an access credential to our cloud computing environment was compromised,” the company’s admitted. “That cloud computing account had not been protected by multifactor authentication. We have now taken steps that include multifactor authentication to secure our authorization and access controls on all accounts.”

All of which leaves users in the same place as usual: with work to do, knowing that if their service providers had done their jobs properly they’d feel a lot safer.

Source: Nostalgic social network ‘Timehop’ loses data from 21 million users