Mozilla: YouTube’s Dislike Button Largely Fails To Stop Unwanted Recommendations

AmiMoJo shares a report from the Mozilla Foundation: YouTube’s user controls — buttons like “Dislike ” and “Not interested” — largely fail to help users avoid unwanted recommendations like misinformation and violent content, according to new research by Mozilla. An accompanying survey also found that YouTube’s controls routinely frustrate and confuse users. Indeed, Mozilla’s research found that people who are experiencing unwanted recommendations and turn to the platform’s user controls for assistance prevent less than half of unwanted recommendations.

This is especially troubling because Mozilla’s past research shows that YouTube recommends videos that violate its very own community guidelines, like misinformation, violent content, hate speech, and spam. For example, one user in this most recent research asked YouTube to stop recommending war footage from Ukraine — but shortly after was recommended even more grisly content from the region. The study, titled “Does This Button Work? Investigating YouTube’s ineffective user controls” is the culmination of months of rigorous qualitative and quantitative research. The study was made possible by the data of more than 20,000 participants who used Mozilla’s RegretsReporter browser extension, and by data about more than 500 million YouTube videos. These are the top findings, as highlighted in the report:
People don’t trust YouTube’s user controls. More than a third (39.3%) of people surveyed felt YouTube’s user controls did not impact their recommendations at all, and 23% felt the controls had a mixed response. Said one interviewee: “Nothing changed. Sometimes I would report things as misleading and spam and the next day it was back in […] Even when you block certain sources they eventually return.”

People take matters into their own hands. Our study found that people did not always understand how YouTube’s controls affect their recommendations, and so took a jury rigged approach instead. People will log out, create new accounts, or use privacy tools just to manage their YouTube recommendations. Said one user: “When the Superbowl came around … if someone recommended a particular commercial, I used to log out of YouTube, watch the commercial, and then log back in.”

The data confirms people are right. The most “effective” user control was “Don’t recommend channel,” but compared to users who do not make use of YouTube’s user controls, only 43% of unwanted recommendations are prevented — and recommendations from the unwanted channel sometimes persist. Other controls were even less effective: The “Not Interested” tool prevented only 11% of unwanted recommendations.

YouTube can fix this problem. YouTube has the power to confront this issue and do a better job at enabling people to control their recommendations. Our research outlines several concrete suggestions to put people back into the driver’s seat, like making YouTube’s controls more proactive, allowing users to shape their own experience; and giving researchers increased access to YouTube’s API and other tools. Further reading: YouTube Targets TikTok With Revenue Sharing For Shorts, Partner Program Expansion

Read more of this story at Slashdot.

Ask.FM Database With 350 Million User Records Allegedly Sold Online

A listing on a popular hacker forum offers 350 million Ask.FM user records for sale in what might be one of the biggest breaches of all time. Cybernews reports: The listing allegedly includes 350 million Ask.FM user records, with the threat actor also offering 607 repositories plus their Gitlab, Jira, and Confluence databases. Ask.FM is a question and answer network launched in June 2010, with over 215 million registered users. The posting also includes a list of repositories, sample git, and sample user data, as well as mentions of the fields in the database: user_id, username, mail, hash, salt, fbid, twitterid, vkid, fbuid, iguid. It appears that Ask.FM is using the weak hashing algorithm SHA1 for passwords, putting them at risk of being cracked and exposed to threat actors.

In response to DataBreaches, the user who posted the database — Data — explained that initial access was gained via a vulnerability in Safety Center. The server was first accessed in 2019, and the database was obtained on 2020-03-14. Data also suggested that Ask.FM knew about the breach as early as back in 2020. While the breach has not been confirmed, the seller called “Data” says he will “vouch all day and night for” listed user data from Ask.FM (ASKfm), the social networking site. “I’m selling the users database of Ask.fm and ask.com,” Data wrote. “For connoisseurs, you can also get 607 repositories plus their Gitlab, Jira, Confluence databases.”

Read more of this story at Slashdot.

Microsoft Won’t Label Fake News As False In An Attempt To Avoid ‘Censorship’ Cries

In an interview with Bloomberg, Microsoft President Brad Smith said the company won’t label social media posts that appear to be false in order to avoid the appearance that the company is trying to censor speech online. From the report: “I don’t think that people want governments to tell them what’s true or false,” Smith said when asked about Microsoft’s role in defining disinformation. “And I don’t think they’re really interested in having tech companies tell them either.” The comments are Smith’s strongest indication yet that Microsoft is taking a unique path to tracking and disrupting digital propaganda efforts.

Smith said Microsoft wanted to provide the public with more information about who is speaking, what they are saying and allow them to come to their own judgment about whether content was true. “We have to be very thoughtful and careful because — and this is also true of every democratic government — fundamentally, people quite rightly want to make up their own mind and they should,” he said. “Our whole approach needs to be to provide people with more information, not less and we cannot trip over and use what others might consider censorship as a tactic.”

Read more of this story at Slashdot.

Charter Must Pay $1.1 Billion After Cable Technician Murdered Customer

Charter Communications must pay over $1.1 billion to the estate and family of an 83-year-old woman murdered in her home by a Spectrum cable technician, a Dallas County Court judge ruled yesterday. Ars Technica reports: A jury in the same court previously ordered Charter to pay $7 billion in punitive damages and $337.5 million in compensatory damages. Judge Juan Renteria lowered the award in a ruling issued yesterday. The damages are split among the estate and four adult children of murder victim Betty Thomas. Renteria did not change the compensatory damages but lowered the punitive damages awarded to the family to $750 million. Pre-judgment interest on the damages pushes Charter’s total liability to over $1.1 billion.

It isn’t surprising that the judge lowered the payout, in which the jury decided punitive damages should be over 20 times higher than what Charter is liable for in compensatory damages. A nine-to-one ratio is often used as a maximum because of a 2003 US Supreme Court ruling that said: “In practice, few awards exceeding a single-digit ratio between punitive and compensatory damages, to a significant degree, will satisfy due process.” Former Spectrum technician Roy Holden pleaded guilty to the 2019 murder of customer Betty Thomas and was sentenced to life in prison in April 2021. Charter was accused of hiring Holden without verifying his employment history and ignoring a series of red flags about his behavior, which included stealing credit cards and checks from elderly female customers.

Read more of this story at Slashdot.

Shutterstock Is Removing AI-Generated Images

Shutterstock appears to be removing images generated by AI systems like DALL-E and Midjourney. Motherboard reports: On Shutterstock, searches for images tagged “Midjourney” yielded several photos with the AI tool’s unmistakable aesthetic, with many having high popularity scores and marked as “frequently used.” But late Monday, the results for “Midjourney” seem to have been reduced, leaving mainly stock photos of the tool’s logo. Other images use tags like “AI generated” — one image, for example, is an illustration of a futuristic building with an image description reading “Ai generated illustration of futuristic Art Deco city, vintage image, retro poster.” The image is part of a collection the artist titled “Midjourney,” which has since been removed from the site. Other images marked “AI generated,” like this burning medieval castle, seem to remain up on the site.

As Ars Technica notes, neither Shutterstock nor Getty Images explicitly prohibits AI-generated images in their terms of service, and Shutterstock users typically make around 15 to 40 percent of what the company makes when it sells an image. Some creators have not taken kindly to this trend, pointing out that these systems use massive datasets of images scraped from the web. […] In other words, the generated works are the result of an algorithmic process which mines original art from the internet without credit or compensation to the original artists. Others have worried about the impacts on independent artists who work for commissions, since the ability for anyone to create custom generated artwork potentially means lost revenue.

Read more of this story at Slashdot.

Denmark and Germany Now Building the World’s Longest Immersed Tunnel

Descending up to 40 meters beneath the Baltic Sea, the world’s longest immersed tunnel will link Denmark and Germany, slashing journey times between the two countries when it opens in 2029. CNN Travel reports: After more than a decade of planning, construction started on the Fehmarnbelt Tunnel in 2020 and in the months since a temporary harbor has been completed on the Danish side. It will host the factory that will soon build the 89 massive concrete sections that will make up the tunnel. “The expectation is that the first production line will be ready around the end of the year, or beginning of next year,” said Henrik Vincentsen, CEO of Femern A/S, the state-owned Danish company in charge of the project. “By the beginning of 2024 we have to be ready to immerse the first tunnel element.”

The tunnel, which will be 18 kilometers (11.1 miles) long, is one of Europe’s largest infrastructure projects, with a construction budget of over 7 billion euros ($7.1 billion). […] It will be built across the Fehmarn Belt, a strait between the German island of Fehmarn and the Danish island of Lolland, and is designed as an alternative to the current ferry service from Rodby and Puttgarden, which carries millions of passengers every year. Where the crossing now takes 45 minutes by ferry, it will take just seven minutes by train and 10 minutes by car. The tunnel, whose official name is Fehmarnbelt Fixed Link, will also be the longest combined road and rail tunnel anywhere in the world. It will comprise two double-lane motorways — separated by a service passageway — and two electrified rail tracks. “Today, if you were to take a train trip from Copenhagen to Hamburg, it would take you around four and a half hours,” says Jens Ole Kaslund, technical director at Femern A/S, the state-owned Danish company in charge of the project. “When the tunnel will be completed, the same journey will take two and a half hours.”

“Today a lot of people fly between the two cities, but in the future it will be better to just take the train,” he adds. The same trip by car will be around an hour faster than today, taking into account time saved by not lining up for the ferry.

Read more of this story at Slashdot.

NYC to Offer Free Broadband to 300,000 Public Housing Residents

New York City is partnering with Charter and Altice to provide free high-speed internet and basic cable TV service to about 300,000 residents of public housing. Bloomberg reports: Called “Big Apple Connect,” the program aims to bridge the digital divide between wealthier residents and lower-income people who lack the tools necessary for remote learning, access to health care and job opportunities, city officials said. An estimated 30% to 40% of people who live in buildings run by the New York City Housing Authority lack broadband, according to the cable providers. The city plans to have the service available in more than 200 NYCHA buildings by the end of 2023.

The program differs from a previous short-term promotion by Altice’s Optimum and Charter’s Spectrum that gave New York City students free internet service after the pandemic hit. Some parents said they were duped into signing up for paid subscriptions after the promotion ended. Under a three-year agreement with the providers, New York will pick up the cost at about $30 per household. The city is in talks with a third major cable TV carrier in the city, Verizon, to join the program. NYCHA residents enrolled in Big Apple Connect will still be able to use the federal Affordable Connectivity Program benefit to save money on their cell phone bills and provide discount of up to $30 per month toward internet and cellular data service, city officials said.

Read more of this story at Slashdot.

Children May Be Losing the Equivalent of One Night’s Sleep a Week From Social Media Use, Study Suggests

Children under 12 may be losing the equivalent of one night’s sleep every week due to excessive social media use, a new study suggests. Insider reports: Almost 70% of the 60 children under 12 surveyed by De Montfort University in Leicester, UK, said they used social media for four hours a day or more. Two thirds said they used social media apps in the two hours before going to bed. The study also found that 12.5% of the children surveyed were waking up in the night to check their notifications.

Psychology lecturer John Shaw, who headed up the study, said children were supposed to sleep for between nine to 11 hours a night, per NHS guidelines, but those surveyed reported sleeping an average of 8.7 hours nightly. He said: “The fear of missing out, which is driven by social media, is directly affecting their sleep. They want to know what their friends are doing, and if you’re not online when something is happening, it means you’re not taking part in it. “And it can be a feedback loop. If you are anxious you are more likely to be on social media, you are more anxious as a result of that. And you’re looking at something, that’s stimulating and delaying sleep.” “TikTok had the most engagement from the children, with 90% of those surveyed saying they used the app,” notes Insider. “Snapchat was used by 84%, while just over half those surveyed said they used Instagram.”

Read more of this story at Slashdot.

When AI Asks Dumb Questions, It Gets Smart Fast

sciencehabit shares a report from Science Magazine: If someone showed you a photo of a crocodile and asked whether it was a bird, you might laugh — and then, if you were patient and kind, help them identify the animal. Such real-world, and sometimes dumb, interactions may be key to helping artificial intelligence learn, according to a new study in which the strategy dramatically improved an AI’s accuracy at interpreting novel images. The approach could help AI researchers more quickly design programs that do everything from diagnose disease to direct robots or other devices around homes on their own.

It’s important to think about how AI presents itself, says Kurt Gray, a social psychologist at the University of North Carolina, Chapel Hill, who has studied human-AI interaction but was not involved in the work. “In this case, you want it to be kind of like a kid, right?” he says. Otherwise, people might think you’re a troll for asking seemingly ridiculous questions. The team “rewarded” its AI for writing intelligible questions: When people actually responded to a query, the system received feedback telling it to adjust its inner workings so as to behave similarly in the future. Over time, the AI implicitly picked up lessons in language and social norms, honing its ability to ask questions that were sensical and easily answerable.

The new AI has several components, some of them neural networks, complex mathematical functions inspired by the brain’s architecture. “There are many moving pieces […] that all need to play together,” Krishna says. One component selected an image on Instagram — say a sunset — and a second asked a question about that image — for example, “Is this photo taken at night?” Additional components extracted facts from reader responses and learned about images from them. Across 8 months and more than 200,000 questions on Instagram, the system’s accuracy at answering questions similar to those it had posed increased 118%, the team reports today in the Proceedings of the National Academy of Sciences. A comparison system that posted questions on Instagram but was not explicitly trained to maximize response rates improved its accuracy only 72%, in part because people more frequently ignored it. The main innovation, Jaques says, was rewarding the system for getting humans to respond, “which is not that crazy from a technical perspective, but very important from a research-direction perspective.” She’s also impressed by the large-scale, real-world deployment on Instagram.

Read more of this story at Slashdot.

GPS Jammers Are Being Used to Hijack Trucks and Down Drones

The world’s freight-carrying trucks and ships use GPS-based satellite tracking and navigation systems, reports ZDNet. But “Criminals are turning to cheap GPS jamming devices to ransack the cargo on roads and at sea, a problem that’s getting worse….”
Jammers work by overpowering GPS signals by emitting a signal at the same frequency, just a bit more powerful than the original. The typical jammers used for cargo hijackings are able to jam frequencies from up to 5 miles away rendering GPS tracking and security apparatuses, such as those used by trucking syndicates, totally useless. In Mexico, jammers are used in some 85% of cargo truck thefts. Statistics are harder to come by in the United States, but there can be little doubt the devices are prevalent and widely used. Russia is currently availing itself of the technology to jam commercial planes in Ukraine.

As we’ve covered, the proliferating commercial drone sector is also prey to attack…. During a light show in Hong Kong in 2018, a jamming device caused 46 drones to fall out of the sky, raising public awareness of the issue.
While the problem is getting worse, the article also notes that companies are developing anti-jamming solutions for drone receivers, “providing protection and increasing the resiliency of GPS devices against jamming attacks.

“By identifying and preventing instances of jamming, fleet operators are able to prevent cargo theft.”

Read more of this story at Slashdot.