The NYT is doing a special podcast on this topic right now.
In one of their episodes, they interview the CEO of YouTube about what they're doing to stop the spread of misinformation on web content platforms like their own.
Her response is that they're no longer tailoring their recommendation models or carousels based purely on engagement alone, but also based on potential harm or impact, because the common misinformation preys on being highly engaging. The biggest example of this is how YouTube is dealing with Covid-19 misinformation, that the "COVID-19 news" carousel on the home page doesn't get much engagement but is important for people to stay informed.
In one of their episodes, they interview the CEO of YouTube about what they're doing to stop the spread of misinformation on web content platforms like their own.
Her response is that they're no longer tailoring their recommendation models or carousels based purely on engagement alone, but also based on potential harm or impact, because the common misinformation preys on being highly engaging. The biggest example of this is how YouTube is dealing with Covid-19 misinformation, that the "COVID-19 news" carousel on the home page doesn't get much engagement but is important for people to stay informed.
It's a good listen if you have the time: https://www.nytimes.com/2020/05/07/podcasts/rabbit-hole-yout...