Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. ×
Google

Is Google's Comment Filtering Tool 'Vanishing' Legitimate Comments? (vortex.com) 96

Slashdot reader Lauren Weinstein writes: Google has announced (with considerable fanfare) public access to their new "Perspective" comment filtering system API, which uses Google's machine learning/AI system to determine which comments on a site shouldn't be displayed due to perceived high spam/toxicity scores. It's a fascinating effort. And if you run a website that supports comments, I urge you not to put this Google service into production, at least for now.

The bottom line is that I view Google's spam detection systems as currently too prone to false positives -- thereby enabling a form of algorithm-driven "censorship" (for lack of a better word in this specific context) -- especially by "lazy" sites that might accept Google's determinations of comment scoring as gospel... as someone who deals with significant numbers of comments filtered by Google every day -- I have nearly 400K followers on Google Plus -- I can tell you with considerable confidence that the problem isn't "spam" comments that are being missed, it's completely legitimate non-spam, non-toxic comments that are inappropriately marked as spam and hidden by Google.

Lauren is also collecting noteworthy experiences for a white paper about "the perceived overall state of Google (and its parent corporation Alphabet, Inc.)" to better understand how internet companies are now impacting our lives in unanticipated ways. He's inviting people to share their recent experiences with "specific Google services (including everything from Search to Gmail to YouTube and beyond), accounts, privacy, security, interactions, legal or copyright issues -- essentially anything positive, negative, or neutral that you are free to impart to me, that you believe might be of interest."
Google

Google Releases an AI Tool For Publishers To Spot and Weed Out Toxic Comments (bbc.com) 195

Google today launched a new technology to help news organizations and online platforms identify and swiftly remove abusive comments on their websites. The technology, called Perspective, will review comments and score them based on how similar they are to comments people said were "toxic" or likely to make them leave a conversation. From a report on BBC: The search giant has developed something called Perspective, which it describes as a technology that uses machine learning to identify problematic comments. The software has been developed by Jigsaw, a division of Google with a mission to tackle online security dangers such as extremism and cyberbullying. The system learns by seeing how thousands of online conversations have been moderated and then scores new comments by assessing how "toxic" they are and whether similar language had led other people to leave conversations. What it's doing is trying to improve the quality of debate and make sure people aren't put off from joining in.
Censorship

CloudFlare Puts Pirate Sites on New IP Addresses, Avoids Cogent Blockade (torrentfreak.com) 88

Earlier this month, several users worldwide reported that they were unable to access pirate websites including the Pirate Bay. It was because the internet backbone network of Cogent Communications had blackholed the CloudFlare IP-address of pirate websites. Less than a week later, CloudFlare is fighting back. From a report on TorrentFreak: The Pirate Bay and dozens of other pirate sites that were blocked by Cogent's Internet backbone are now accessible again. CloudFlare appears to have moved the sites in question to a new pair of IP-addresses, effectively bypassing Cogent's blackhole. [...] As of yesterday, the sites in question have been assigned the IP-addresses 104.31.16.3 and 104.31.17.3, still grouped together. Most, if not all of the sites, are blocked by court order in the UK so this is presumably done to prevent ISP overblocking of 'regular' CloudFlare subscribers.
Twitter

Twitter Announces (More) Hate-Speech Fighting Tools (Again) (cnn.com) 341

Building on anti-harassment tools announced in November, Twitter is now "trying to shake its reputation as a haven for online harassment" with still more new internal algorithms and features, reports CNN. An anonymous reader quotes their report: The changes include preventing serial abusers from creating new accounts, a new "safe search" function and blocking potentially abusive and "low-quality" tweets from appearing in conversations, Twitter's engineering chief Ed Ho said in a blog post. Twitter is working on identifying users that have been permanently suspended and prevent them from creating new accounts, Ho said. This new measure specifically targets "accounts that are created only to abuse and harass others," he said, a problem that has long plagued the platform.

The new safe search function prevents tweets that are abusive, or from blocked and muted accounts, from appearing in users' search results. Those tweets can still be found if people want to see them, but they "won't clutter search results any longer," Ho said. And Twitter will now collapse tweet replies that are potentially abusive or low quality -- like duplicate tweets or content that appears to be automated. But those tweets "will still be accessible to those who seek them out," Ho said.

The blog post announces Twitter's ultimate goal is "a significant impact that people can feel," arguing that freedom of speech for all viewpoints is "put in jeopardy when abuse and harassment stifle and silence those voices."
Piracy

Internet Backbone Provider Cogent Blocks Pirate Bay and Other 'Pirate' Sites (torrentfreak.com) 186

Several Pirate Bay users from ISPs all over the world have been unable to access their favorite torrent site for more than a week. Their requests are being stopped in the Internet backbone network of Cogent Communications, which has blackholed the CloudFlare IP-address of The Pirate Bay and many other torrent and streaming sites, reports TorrentFreak. From the article: When the average Internet user types in a domain name, a request is sent through a series of networks before it finally reaches the server of the website. This also applies to The Pirate Bay and other pirate sites such as Primewire, Movie4k, TorrentProject and TorrentButler. However, for more than a week now the US-based backbone provider Cogent has stopped passing on traffic to these sites. The sites in question all use CloudFlare, which assigned them the public IP-addresses 104.31.18.30 and 104.31.19.30. While this can be reached just fine by most people, users attempting to pass requests through Cogent's network are unable to access them.
Censorship

Tor's Ooniprobe, Now Available On Android and iOS, Helps People Track Internet Censorship (cnn.com) 27

In 2012, researchers at Tor announced Ooniprobe, an open-source tool to collect data about local meddling with the computer's network connections, and also whether the government was censoring something. The team has now released a new app, available for Android and iOS, which makes it easier than ever to tell what your government is up to on the web. From a report on CNN Money: The Open Observatory of Network Interference (OONI), which monitors networks for censorship and surveillance, is launching Ooniprobe, a mobile app to test network connectivity and let you know when a website is censored in your area. The app tests over 1,200 websites, including Facebook, Twitter, and WhatsApp. You can decide how long to run the test, but the default is 90 seconds and would test between 10 and 20 websites depending on bandwidth. Links to blocked websites are listed in red, while available sites are green. Service providers, sometimes controlled by the government, don't always shutdown the internet entirely -- for instance, Facebook.com might be inaccessible while CNN.com still works. "Not only we will be able to gather more data and more evidence, but we will be able to engage and bring the issue of censorship to the attention of more people," Arturo Filasto, chief developer for the Ooniprobe app, told CNNTech.
Censorship

Why Has Cameroon Blocked the Internet? (bbc.com) 87

It has been over three weeks since English-speaking parts of Cameroon, a country on the west coast of Africa between Nigeria and Gabon, has had no internet connectivity. Residents believe, according to a BBC report, that the government is behind it. From the report: The two regions affected, South-West and North-West, have seen anti-government protests in recent months. Just a day before services disappeared, the Ministry of Posts and Telecommunications issued a statement in which it warned social media users of criminal penalties if they were to "issue or spread information, including by way of electronic communications or information technology systems, without any evidence." There has been no official comment about the internet since then (or any credible reports of technical faults) leading many Cameroonians to conclude that the severing of services is part of government attempts to stifle dissent. In criticising their government, some Cameroonians have also taken aim at the mobile phone companies who provide the services through which many access the internet. These firms may not have been able to prevent the outage, since they all rely on fibre-optic infrastructure provided by a state-owned company, but nor have they been objecting publicly about the interruption to their services.
Desktops (Apple)

Apple Seemingly Censors UltraFine 5K Monitor Reviews After Poor Feedback (thenextweb.com) 97

It appears Apple is filtering and censoring bad reviews of the LG's UltraFine 5K display. From a report on The Next Web: The deletion was first spotted by a Reddit user four days ago. Though it's possible the reviews were removed for some other reason, at first glance, it looks like censorship. It's not a good look for the company. Apple said it was getting out of the monitor business, and instead chose to work more closely with third-party partners, heavily featuring LG's 5K and 4K UltraFine displays at its recent MacBook Pro unveiling. But then the monitor received multiple negative reviews from users who were experiencing issues such as the screen failing to wake up from sleep. The Reddit post also points out that: "In many cases, attempts to fix the problem through physical reconnection[sic] of the monitor, or manual restarts, have caused the attached Mac to crash, become otherwise unresponsive, or develop problems with the touch bar (where equipped)."
Security

ProtonMail Adds Tor Onion Site To Fight Risk Of State Censorship (techcrunch.com) 26

ProtonMail now has a home on the dark web. The encrypted email provider announced Thursday it will allow its users to access the site through the Tor anonymity service. From a report: Swiss-based PGP end-to-end encrypted email provider, ProtonMail, now has an onion address, allowing users to access its service via a direct connection to the Tor anonymizing network -- in what it describes as an active measure aimed at defending against state-sponsored censorship. The startup, which has amassed more than two million users for its e2e encrypted email service so far, launching out of beta just over a year ago, says it's worried about an increased risk of state-level blocking of pro-privacy tools -- pointing to recent moves such as encryption messaging app Signal being blocked in Egypt, and the UK passing expansive surveillance legislation that mandates tracking of web activity and can also require companies to eschew e2e encryption and backdoor products. The service also saw a bump in sign ups after the election of Donald Trump as US president, last fall -- with web users apparently seeking a non-US based secure email provider in light of the incoming commander-in-chief's expansive digital surveillance powers.
The Internet

Thousands Of Cubans Now Have Internet Access (ap.org) 70

There's been a dramatic change in one of the world's least-connected countries. An anonymous reader quotes the AP: Since the summer of 2015, the Cuban government has opened 240 public Wi-Fi spots in parks and on street corners across the country... The government estimates that 100,000 Cubans connect to the internet daily. A new feature of urban life in Cuba is the sight of people sitting at all hours on street corners or park benches, their faces illuminated by the screen of smartphones connected by applications such as Facebook Messenger to relatives in Miami, Ecuador or other outposts of the Cuban diaspora...

Cuban ingenuity has spread internet far beyond those public places: thousands of people grab the public signals through commercially available repeaters, imported illegally into Cuba and often sold for about $100 -- double the original price. Mounted on rooftops, the repeaters grab the public signals and create a form of home internet increasingly available in private rentals for tourists and cafes and restaurants for Cubans and visitors alike.

The article also points out that last month, for the first time ever, 2,000 Cubans began receiving home internet access.
Censorship

Russia Demands LinkedIn App Takedown, Apple and Google Comply (fortune.com) 110

Russia has forced Apple and Google to remove the LinkedIn mobile app from their Russian application markets, the latest chapter in a months-long campaign against the professional networking site. From a report on Fortune: A recently-passed Russian law requires that any company holding data on Russians house that data within Russia. Russia began blocking LinkedIn's website last November under that law, which some critics argue is an indirect form of censorship. The removal of the LinkedIn app from Apples App Store and Google's Play shows the willingness of major internet gatekeepers to comply with individual nations' data-control laws, on both the web and mobile devices.
China

Apple Removes NYTimes App in China, Shows How Far It Is Willing To Go To Please Local Authority (theguardian.com) 174

Apple has removed the New York Times app from its store in China after a government request, in an example of how far the company will go to please the authorities in its third-largest market. From a report: China operates what is thought to be the largest internet censorship regime in the world, blocking thousands of foreign websites viewed as a threat by the ruling Communist party. Google, Twitter, Facebook Youtube and Instagram are all inaccessible. Apple removed the English and Chinese-language versions of the New York Times app on 23 December, although it was not immediately clear why. "We have been informed that the app is in violation of local regulations," said Carolyn Wu, an Apple spokeswoman. "As a result the app must be taken down off the China app store. When this situation changes the app store will once again offer the New York Times app for download in China."
Facebook

Facebook Is Sorry for Taking Down a Photo of a Nude Neptune Statue (fortune.com) 159

Facebook has apologized for mistakenly blocking a photo of a famous statue for being "sexually explicit." From a report on Fortune: The social media giant flagged a photograph of a 16th-century statue of the sea god Neptune in the Italian city of Bologna. The picture of the sculpture -- which was created in the 1560s -- was featured on the Facebook page of local writer and art historian Elisa Barbari called "Stories, curiosities and view of Bologna." Facebook told Barbari that the picture violated the company's privacy policies. "It shows an image with content that is explicitly sexual and which excessively shows the body or unnecessarily concentrates on body parts," the company said in a statement. The company added: "The usage of images or video of nude bodies or plunging necklines is not allowed, even if the use is for artistic or educational reasons." Facebook later said that blocking the photo was a mistake.
Facebook

Germany Considers Fining Facebook $522,000 Per Fake News Item (heatst.com) 333

"The government of Germany is considering imposing a legal regime that would allow fining social networks such as Facebook up to 500,000 euros ($522,000) for each day the platform leaves a 'fake news' story up without deleting it," according to a story shared by schwit1. PC Magazine has more details: The law would reportedly apply to other social networks as well. "If after the relevant checks Facebook does not immediately, within 24 hours, delete the offending post then [it] must reckon with severe penalties of up to 500,000 euros," Germany's parliamentary chief of the Social Democrat party Thomas Oppermann said in an interview with Germany's Der Spiegel magazine, according to a report from Heat Street. Under the law, "official and private complainants" would be able to flag news on Facebook as fake, Heat Street reported. Facebook and other affected social networks would have to create "in-country offices focused on responding to takedown demands," the report says. The bill, slated for consideration next year, is said to have bipartisan support. According to the article, "Lawmakers in the country are reportedly hoping it will prevent Russia from interfering in Germany's elections next year."
Electronic Frontier Foundation

2016 Saw A Massive Increase In Encrypted Web Traffic (eff.org) 91

EFF's "Deeplinks" blog has published nearly two dozen "2016 in Review" posts over the last nine days, one of which applauds 2016 as "a great year for adoption of HTTPS encryption for secure connections to websites." An anonymous reader writes: In 2016 most pages viewed on the web were encrypted. And over 21 million web sites obtained security certificates -- often for the first time -- through Let's Encrypt. But "a sizeable part of the growth in HTTPS came from very large hosting providers that decided to make HTTPS a default for sites that they host, including OVH, Wordpress.com, Shopify, Tumblr, Squarespace, and many others," EFF writes. Other factors included the support of Transport Layer Security (TLS) 1.3 by Firefox, Chrome, and Opera.
Other "2016 in Review" posts from EFF include Protecting Net Neutrality and the Open Internet and DRM vs. Civil Liberties. Click through for a complete list of all EFF "2016 in Review" posts.
Government

Governments Shut Down the Internet More Than 50 Times in 2016 (thewire.in) 34

An anonymous reader writes: Governments around the world shut down the internet more than 50 times in 2016 -- suppressing elections, slowing economies and limiting free speech. In the worst cases internet shutdowns have been associated with human rights violations, Deji Olukotun, senior global advocacy manager at digital rights organisation Access Now told IPS. "What we have found is that internet shutdowns go hand in hand with atrocities," said Olukotun. "In Ethiopia there's been consistent blocking this year of social media and internet." Dozens of people have died in protests in Ethiopia in 2016, "many of them during the kind of blackout where it's difficult to report on what's happening," he said. Several leaders used internet shutdowns to affect democratic processes, including elections. "In Uganda in February 2016 there was a shutdown of social media networks by President Museveni and that again happened in Gambia (in December) surrounding the election," Olukotun added. In other cases, three governments chose to shut down the internet because they thought that it would stop students from cheating on their exams, he said. "On the whole most governments want to expand internet access," said Olukotun. However governments do not seem to have taken into account the potential repercussions of the shutdowns, beyond the limits of free speech. According to an estimate, internet shutdowns resulted in a loss of $2.4 billion in 2015.
Encryption

Encrypted Messaging App Signal Uses Google To Bypass Censorship (pcworld.com) 87

Developers of the popular Signal secure messaging app have started to use Google's domain as a front to hide traffic to their service and to sidestep blocking attempts. Bypassing online censorship in countries where internet access is controlled by the government can be very hard for users. It typically requires the use of virtual private networking (VPN) services or complex solutions like Tor, which can be banned too. From a report on PCWorld: Open Whisper Systems, the company that develops Signal -- a free, open-source app -- faced this problem recently when access to its service started being censored in Egypt and the United Arab Emirates. Some users reported that VPNs, Apple's FaceTime and other voice-over-IP apps were also being blocked. The solution from Signal's developers was to implement a censorship circumvention technique known as domain fronting that was described in a 2015 paper by researchers from University of California, Berkeley, the Brave New Software project and Psiphon. The technique involves sending requests to a "front domain" and using the HTTP Host header to trigger a redirect to a different domain. If done over HTTPS, such redirection would be invisible to someone monitoring the traffic, because the HTTP Host header is sent after the HTTPS connection is negotiated and is therefore part of the encrypted traffic.
Social Networks

Turkey Blocks Tor's Anonymity Network (engadget.com) 74

An anonymous reader writes: Turkey's President Erdogan and the ruling AKP party are increasingly bent on silencing online dissent, and that now affects you even if you're smart enough to evade typical censorship methods. Watchdog group Turkey Blocks has confirmed that Turkey is blocking the Tor anonymity network's direct access mode for most users. You can still use a bridge mode for now, but there are hints that internet providers might be hurting performance even then. The restrictions come alongside a recent government ban on virtual private network services.
China

China Takes Action On Thousands Of Websites For 'Harmful', Obscene Content (reuters.com) 60

China has shut down or "dealt with" thousands of websites for sharing "harmful" erotic or obscene content since April, the state's office for combating pornography and illegal publications announced on Thursday. From a report on Reuters: The office said 2,500 websites were prosecuted or shut down and more than 3 million "harmful" posts were deleted in eight months up to December during a drive to "purify" the internet in China and protect youth, the official Xinhua news agency reported. The government has tightening its grip on Chinese cyberspace in recent months, in particular placing new restrictions on the fast-growing live-streaming industry. The state has a zero-tolerance approach to what it considers lewd, smutty or illegal content and has in past crackdowns removed tens of thousands of websites in a single year.
Twitter

Twitter Reinstates White Nationalist Leader's Account (buzzfeed.com) 470

An anonymous reader quotes BuzzFeed: On Saturday evening, Twitter reinstated -- with verification -- the account of Richard Spencer, a leading figure of the so-called alt-right movement, and the head of the white nationalist think tank, The National Policy Institute. Spencer's account was suspended mid-November as part of a larger cull of prominent alt-right accounts... However, according to Twitter, Spencer was banned on a technicality: creating multiple accounts with overlapping uses. Twitter's multiple account policy was put in place as a safeguard to help curb dog piling and targeted harassment. [Twitter] offered to reinstate one of Spencer's accounts if he agreed to follow the company's protocols.
Vox says the move "raises the question of to what extent Twitter intends to enforce the 'hateful conduct' policy." But the suspension had also been criticized by David Frum, a senior editor at the Atlantic, who wrote that "The culture of offense-taking, platform-denying, and heckler-vetoing...lets loudmouths and thugs present themselves as heroes of free thought. They do not deserve this opportunity... today, a neo-Nazi has more right to build an arsenal of weapons and drill a militia than to speak on Twitter." But BuzzFeed points out that though the account's been reinstated, Spencer "is now tip-toeing around the company's three strike policy, which carries a permanent suspension."

Slashdot Top Deals