Group items tagged

YouTube has inadvertently created a dangerous on-ramp to extremism by combining two things: a business model that rewards provocative videos with exposure and advertising dollars, and an algorithm that guides users down personalized paths meant to keep them glued to their screens

“If I’m YouTube and I want you to watch more, I’m always going to steer you toward Crazytown.”

94 percent of Americans ages 18 to 24 use YouTube, a higher percentage than for any other online service

YouTube has been a godsend for hyper-partisans on all sides. It has allowed them to bypass traditional gatekeepers and broadcast their views to mainstream audiences, and has helped once-obscure commentators build lucrative media businesses

Bellingcat, an investigative news site, analyzed messages from far-right chat rooms and found that YouTube was cited as the most frequent cause of members’ “red-pilling” — an internet slang term for converting to far-right beliefs

The internet was an escape. Mr. Cain grew up in postindustrial Appalachia and was raised by his conservative Christian grandparents. He was smart, but shy and socially awkward, and he carved out an identity during high school as a countercultural punk. He went to community college, but dropped out after three semesters. Broke and depressed, he resolved to get his act together. He began looking for help in the same place he looked for everything: YouTube.

they rallied around issues like free speech and antifeminism, portraying themselves as truth-telling rebels doing battle against humorless “social justice warriors.” Their videos felt like episodes in a long-running soap opera, with a constant stream of new heroes and villains. To Mr. Cain, all of this felt like forbidden knowledge — as if, just by watching some YouTube videos, he had been let into an exclusive club. “When I found this stuff, I felt like I was chasing uncomfortable truths,” he told me. “I felt like it was giving me power and respect and authority.”

YouTube’s executives announced that the recommendation algorithm would give more weight to watch time, rather than views. That way, creators would be encouraged to make videos that users would finish, users would be more satisfied and YouTube would be able to show them more ads.

A month after its algorithm tweak, YouTube changed its rules to allow all video creators to run ads alongside their videos and earn a portion of the revenue they generated.

Many right-wing creators already made long video essays, or posted video versions of their podcasts. Their inflammatory messages were more engaging than milder fare. And now that they could earn money from their videos, they had a financial incentive to churn out as much material as possible.

Several current and former YouTube employees, who would speak only on the condition of anonymity because they had signed confidentiality agreements, said company leaders were obsessed with increasing engagement during those years. The executives, the people said, rarely considered whether the company’s algorithms were fueling the spread of extreme and hateful political content.

Google Brain’s researchers wondered if they could keep YouTube users engaged for longer by steering them into different parts of YouTube, rather than feeding their existing interests. And they began testing a new algorithm that incorporated a different type of A.I., called reinforcement learning. The new A.I., known as Reinforce, was a kind of long-term addiction machine. It was designed to maximize users’ engagement over time by predicting which recommendations would expand their tastes and get them to watch not just one more video but many more.

YouTube’s recommendations system is not set in stone. The company makes many small changes every year, and has already introduced a version of its algorithm that is switched on after major news events to promote videos from “authoritative sources” over conspiracy theories and partisan content. This past week, the company announced that it would expand that approach, so that a person who had watched a series of conspiracy theory videos would be nudged toward videos from more authoritative news sources. It also said that a January change to its algorithm to reduce the spread of so-called “borderline” videos had resulted in significantly less traffic to those videos.

the bulk of his media diet came from far-right channels. And after the election, he began exploring a part of YouTube with a darker, more radical group of creators. These people didn’t couch their racist and anti-Semitic views in sarcastic memes, and they didn’t speak in dog whistles. One channel run by Jared Taylor, the editor of the white nationalist magazine American Renaissance, posted videos with titles like “‘Refugee’ Invasion Is European Suicide.” Others posted clips of interviews with white supremacists like Richard Spencer and David Duke.

As Mr. Molyneux promoted white nationalists, his YouTube channel kept growing. He now has more than 900,000 subscribers, and his videos have been watched nearly 300 million times. Last year, he and Ms. Southern — Mr. Cain’s “fashy bae” — went on a joint speaking tour in Australia and New Zealand, where they criticized Islam and discussed what they saw as the dangers of nonwhite immigration. In March, after a white nationalist gunman killed 50 Muslims in a pair of mosques in Christchurch, New Zealand, Mr. Molyneux and Ms. Southern distanced themselves from the violence, calling the killer a left-wing “eco-terrorist” and saying that linking the shooting to far-right speech was “utter insanity.” Neither Mr. Molyneux nor Ms. Southern replied to a request for comment. The day after my request, Mr. Molyneux uploaded a video titled “An Open Letter to Corporate Reporters,” in which he denied promoting hatred or violence and said labeling him an extremist was “just a way of slandering ideas without having to engage with the content of those ideas.”

Unlike most progressives Mr. Cain had seen take on the right, Mr. Bonnell and Ms. Wynn were funny and engaging. They spoke the native language of YouTube, and they didn’t get outraged by far-right ideas. Instead, they rolled their eyes at them, and made them seem shallow and unsophisticated.

“I noticed that right-wing people were taking these old-fashioned, knee-jerk, reactionary politics and packing them as edgy punk rock,” Ms. Wynn told me. “One of my goals was to take the excitement out of it.”

Ms. Wynn and Mr. Bonnell are part of a new group of YouTubers who are trying to build a counterweight to YouTube’s far-right flank. This group calls itself BreadTube, a reference to the left-wing anarchist Peter Kropotkin’s 1892 book, “The Conquest of Bread.” It also includes people like Oliver Thorn, a British philosopher who hosts the channel PhilosophyTube, where he posts videos about topics like transphobia, racism and Marxist economics.

The core of BreadTube’s strategy is a kind of algorithmic hijacking. By talking about many of the same topics that far-right creators do — and, in some cases, by responding directly to their videos — left-wing YouTubers are able to get their videos recommended to the same audience.

What is most surprising about Mr. Cain’s new life, on the surface, is how similar it feels to his old one. He still watches dozens of YouTube videos every day and hangs on the words of his favorite creators. It is still difficult, at times, to tell where the YouTube algorithm stops and his personality begins.

It’s possible that vulnerable young men like Mr. Cain will drift away from radical groups as they grow up and find stability elsewhere. It’s also possible that this kind of whiplash polarization is here to stay as political factions gain and lose traction online.

I’ve learned now that you can’t go to YouTube and think that you’re getting some kind of education, because you’re not.

Many Arab LGBT+ activists - even those who work closely with Facebook - say the rules are still poorly enforced. In recent months, LGBT+ groups have formed their own Facebook groups to mass-report viral hate speech, appealing to the company to act quickly. In June, Facebook set up phone calls with prominent Arab LGBT+ activists and groups to hear their concerns, according to four participants in those calls. But activists said they were unconvinced that moderators were well-trained to identify all Arabic hate speech.

The ANKH survey found the problem persisted into August, with 77% of respondents saying they were subjected to online hate speech.

"People were sending them to my family and colleagues asking if it was me," he said. "I never expected things would turn real."

Because the term “AI” is ambiguous, it makes having clear discussions more difficult. In one sense, it is the name of a subfield of computer science. In another, it can refer to the computing techniques developed in that subfield, most of which are now focused on pattern matching based on large data sets and the generation of new media based on those patterns. Finally, in marketing copy and start-up pitch decks, the term “AI” serves as magic fairy dust that will supercharge your business.

output can seem so plausible that without a clear indication of its synthetic origins, it becomes a noxious and insidious pollutant of our information ecosystem

the people selling this technology propose that text synthesis machines could fix various holes in our social fabric: the lack of teachers in K–12 education, the inaccessibility of health care for low-income people and the dearth of legal aid for people who cannot afford lawyers, just to name a few

Not only do we risk mistaking synthetic text for reliable information, but also that noninformation reflects and amplifies the biases encoded in its training data—in this case, every kind of bigotry exhibited on the Internet. Moreover the synthetic text sounds authoritative despite its lack of citations back to real sources. The longer this synthetic text spill continues, the worse off we are, because it gets harder to find trustworthy sources and harder to trust them when we do.

the systems rely on enormous amounts of training data that are stolen without compensation from the artists and authors who created it in the first place

the task of labeling data to create “guardrails” that are intended to prevent an AI system’s most toxic output from seeping out is repetitive and often traumatic labor carried out by gig workers and contractors, people locked in a global race to the bottom for pay and working conditions.

employers are looking to cut costs by leveraging automation, laying off people from previously stable jobs and then hiring them back as lower-paid workers to correct the output of the automated systems. This can be seen most clearly in the current actors’ and writers’ strikes in Hollywood, where grotesquely overpaid moguls scheme to buy eternal rights to use AI replacements of actors for the price of a day’s work and, on a gig basis, hire writers piecemeal to revise the incoherent scripts churned out by AI.

too many AI publications come from corporate labs or from academic groups that receive disproportionate industry funding. Much is junk science—it is nonreproducible, hides behind trade secrecy, is full of hype and uses evaluation methods that lack construct validity

We urge policymakers to instead draw on solid scholarship that investigates the harms and risks of AI—and the harms caused by delegating authority to automated systems, which include the unregulated accumulation of data and computing power, climate costs of model training and inference, damage to the welfare state and the disempowerment of the poor, as well as the intensification of policing against Black and Indigenous families. Solid research in this domain—including social science and theory building—and solid policy based on that research will keep the focus on the people hurt by this technology.

President Joe Biden condemned the Islamophobic attack that led to the death of the Palestinian child earlier this weekend, as he should have: but he, and all of us, have to go after the background music that gives fuel to that kind of atrocity, and worse.