Filter bubble

Coined by Eli Pariser in 2011 about search and social feeds, and it applies to any recommendation system that learns from behaviour. Watch three videos on one subject and the homepage fills with more of the same, so the range of what you are offered shrinks. For creators it cuts both ways. Reaching an audience that already watches you gets easier while breaking into a new one gets harder, because the people who might like your work are being served something else. Treating it as deliberate censorship misreads it. It falls out of a system optimising for what somebody clicked yesterday rather than out of any editorial decision.