Context: "This dataset represents a small subset of AI-generated fake podcasts that were flagged during attempts to add ... The goal of sharing this dataset is to ... combat spam."
TheChaplain · 3h ago
Future elections will be interesting.
It's obvious with AI/LLM these days you can very quickly create wast amounts of information that drowns out whatever is on the agenda.
bananalychee · 2h ago
We've been there for a while. Most people get their news from social media networks that persistently funnel users into content silos that validate and reinforce their beliefs, and even traditional news outlets fall prey to biased and forced narratives all the time. Paid propagandists and bot farms have been adding noise for several years. You have to take everything with a grain of salt unless it's verifiable through citations from primary sources or references to data sources for statistical analyses, which are rare to find on news websites (although they are better at filtering out false rumors) and even rarer on popular social media networks that favor short-form text content. I feel like skepticism and critical thinking are trending in response, but that might just be my filter bubble.
mindctrl-org · 4h ago
How long before dead internet theory is for real for real
TimedToasts · 3h ago
Negative several years.
sram1337 · 3h ago
I take issue with the term "fake podcast." This is like calling AI generated art "fake art."
The issue they have is with low-quality content. If the AI generated content was better than most human-created podcasts and were making their engagement numbers go up, I doubt they would be calling them fake or removing them.
egypturnash · 3h ago
It's fake art made by pouring a bunch of images into an algorithmic hopper and spitting out something vaguely like them without paying a single cent to anyone whose data was grabbed for this abuse of fair use, at a huge cost of power hidden away from the user.
sram1337 · 1h ago
I disagree its "fake art."
Criticize the process of creating it all you want.
It's obvious with AI/LLM these days you can very quickly create wast amounts of information that drowns out whatever is on the agenda.
The issue they have is with low-quality content. If the AI generated content was better than most human-created podcasts and were making their engagement numbers go up, I doubt they would be calling them fake or removing them.
Criticize the process of creating it all you want.