This week saw the release of the long awaited final season of Stranger Things. My younger daughter and I are still working our way through the previous seasons. We should have started earlier but, c’est la vie, we won’t be ready to start the new season for at least another week (I’m traveling for work which makes it worse…)
Yesterday when we sat down to watch a couple of more episodes of season 3 my daughter mentioned she had seen a minor spoiler for season 5 on Tik Tok or Instagram or whatever. She’s a teenager and she spend too much time doom scrolling, especially during the school holiday, which for Singaporean kids started recently.
This got me thinking. This sort of spoiler filled social media has been a thing for a while. I think it hit its height during the release of Avengers: Endgame. People want to talk about the shows or things they like (to say nothing of people who will do anything to make a living via social media…) and the algorithm is going to push that into your feed if it’s popular and you have ever interacted with any content it thinks is similar. So if you have ever scrolled through Stranger Things memes you’re going to get Stranger Things season 5 spoilers.
AI should fix this. The social media companies should have a function where I can tell their AI “I don’t want any Stranger Things spoilers”. And the AI should then filter all that shit out of my feed until I tell it otherwise. The sort of fuzzy logic language and image processing that GenAI is lauded for should be able to handle this.
In fact, this would be a good feature to clean up your feed about anything. The algorithms are too damn aggressive. Someone sends you one funny clip about something you don’t care about for a laugh and your feed is polluted with that shit for weeks as the algorithm begs you to interact with more of the same.
I have not found the “generate a unique emoji” or search with AI features of AI in social media apps to be of any utility. I think the ability to craft my feed through instructions to the AI would be much more useful.
Maybe they don’t want us to filter things out, I’m not sure if there is anyway to properly filter your feed, I’m too lazy to find out, the more bullshit in my feed the less I use social media. Which is a good thing. But maybe the issue is the AI is not really so powerful?
I’ve been frustrated with the AI they shoved into search in the past. I have not been able to find specific posts that include images or language I describe to the search AI with any success so maybe the AI is not so good. Several times I’ve tried to use the Meta AI in search to find a reel in Instagram where a specific creator talked about a specific topic so I can share it. It always fails. It can find the creator but doesn’t seem to filter the feed at all. It prompts for follow up questions but can’t seem to filter the first set of results for the input I give it. I’ve tried a first search for a creator and check, got them. Then, when I ask it to find the specific post where said creator talked about topic X? Nope, I get random fucking results about X and never the one from the creator I just searched for.
So maybe the AIs are not as good as they want us to think. Or maybe they just need to tune them. But I think a spoiler blocker, or a more general topic blocker, feature using AI would be very useful.