As disinformation and hate thrive online, YouTube quietly changed how it moderates content

Nick Logan · CBC News

· Posted: Jun 14, 2025 4:00 AM EDT | Last Updated: 8 hours ago

YouTube, the world's largest video platform, appears to person changed its moderation policies to let much contented that violates its ain rules to stay online.

The alteration happened softly successful December, according to The New York Times, which reviewed grooming documents for moderators indicating that a video could enactment online if the offending worldly did not relationship for much than 50 per cent of the video's duration — that's treble what it was anterior to the caller guidelines.  

YouTube, which sees 20 cardinal videos uploaded a day, says it updates its guidance regularly and that it has a "long-standing signifier of applying exceptions" erstwhile it suits the nationalist involvement oregon erstwhile thing is presented successful an educational, documentary, technological oregon creator context.

"These exceptions use to a tiny fraction of the videos connected YouTube, but are captious for ensuring important contented remains available," YouTube spokesperson Nicole Bell said successful a connection to CBC News this week.

But successful a clip erstwhile societal media platforms are awash with misinformation and conspiracy theories, determination are concerns that YouTube is lone opening the doorway for much radical to spread problematic oregon harmful contented — and to marque a nett doing so. 

YouTube isn't alone. Meta, which owns Facebook and Instagram, dialled backmost its contented moderation earlier this year, and Elon Musk sacked Twitter's moderators erstwhile helium purchased the level successful 2022 and rebranded it arsenic X. 

"We're seeing a contention to the bottommost now," Imran Ahmed, CEO for the U.S.-based Center for Countering Digital Hate, told CBC News. "What we're going to spot is simply a maturation successful the system astir hatred and disinformation."

WATCH | Experts pass Meta's moderation move volition apt summation misinformation: 

Public involvement vs nationalist harm

YouTube's extremity is "to support escaped expression," Brooks said successful her statement, explaining that easing its assemblage guidelines "reflect the caller types of content" connected the platform. 

For example, she said, a long-form podcast containing one abbreviated clip of unit whitethorn nary longer request to be removed. 

The Times reported Monday that examples presented to YouTube staff included a video successful which idiosyncratic utilized a derogatory word for transgender people during a treatment astir hearings for U.S. President Donald Trump's furniture appointees, and different that shared mendacious accusation astir COVID-19 vaccines but that did not outright archer radical not to get vaccinated.

A level similar YouTube does person to marque immoderate "genuinely precise hard decisions" erstwhile moderating content, says Matt Hatfield, enforcement manager of the Canadian integer rights radical OpenMedia.

LISTEN |  How Canada has travel to play an outsized relation successful far-right misinformation:

He believes platforms bash instrumentality the issue seriously, but helium says there's a equilibrium betwixt removing harmful oregon amerciable content, specified arsenic kid maltreatment worldly oregon wide incitements to violence, and allowing contented to enactment online, even if it's violative to many or contains immoderate mendacious information. 

The problem, he says, is that societal media platforms besides "create environments that promote immoderate atrocious behaviour" among creators, who similar to locomotion the enactment of what's acceptable.

"The halfway exemplary of these platforms is to support you clicking, support you watching, get you to effort a video from idiosyncratic you've ne'er experienced earlier and past instrumentality with that person."

And that's what concerns Ahmed.

He says these companies put profits implicit online information and that they don't look consequences because there are nary regulations forcing them to bounds what tin beryllium posted connected their platforms.

He believes YouTube's relaxed policies will lone promote much radical to exploit them.

How good YouTube is moderating 

In a caller transparency report, YouTube said it had removed astir 2.9 cardinal channels containing much than 47 cardinal videos for assemblage guideline violations successful the archetypal quarter — that came after the reported argumentation change. 

The overwhelming bulk of those, 81.8 per cent, were considered spam, but different reasons included violence, hateful oregon abusive worldly and kid safety.

LISTEN | Why you're being tormented by ads algorithms and AI slop:

Hatfield says determination is simply a nationalist involvement successful having harmful contented similar that removed, but that doesn't mean each arguable oregon violative contented indispensable go.

However, helium says YouTube does marque mistakes successful contented moderation, explaining that it judges idiosyncratic videos successful a sort of "vacuum" without considering however each portion of content fits into a broader context.

"Some contented can't truly beryllium reasonably interpreted successful that way."

Regulations not a cleanable solution

Ahmed says companies should beryllium held accountable for the contented connected their platforms done authorities regulation.

He pointed to Canada's arguable but now-scuttled Online Harms Act, besides known arsenic Bill C-63, arsenic an example. It projected heavier sentences, caller regulatory bodies and changes to a fig of laws to tackle online abuse. The measure died erstwhile erstwhile premier curate Justin Trudeau announced his resignation and prorogued Parliament backmost successful January. Ahmed says helium hopes the caller government under Prime Minister Mark Carney will enact akin legislation. 

Hatfield says helium liked parts of that act, but his radical yet opposed it after it tacked connected immoderate different changes to the Criminal Code and Human Rights Act that helium says were unrelated to the platforms.

He says groups similar OpenMedia would person liked to spot a strategy addressing concern models that promote users to station and nett disconnected of "lawful but awful" content. 

"We're not going to person a hate-free internet," helium said. "We tin person an net that makes it little profitable to dispersed definite types of hatred and misinformation." 

WATCH | How radical tin go much discerning quality consumers: