- Jan 29, 2025
- 870
- 209
During debates and interviews, politicians can make dozens of bold claims.
Normally, journalists fact-check these statements, but their analyses often appear hours, or even days later, after millions of people have already watched and discussed the speech.
One student decided to change that by creating a Chrome extension called InTruth.
It runs directly on top of videos and fact-checks politicians' statements in real time.
First, the AI converts speech into text, then identifies claims that can actually be verified.
It searches for information from reliable sources, analyzes the findings using Claude, and displays the results almost instantly.
Instead of simply labeling a statement as "true" or "false,"
InTruth uses several categories: "True," "Mostly True," "Misleading," "False," and "Can't Be Verified."
Each verdict is accompanied by links to the sources, allowing users to review the evidence and draw their own conclusions.
What once required an entire team of journalists and fact-checkers can now be done by a single student using modern AI models.
It's another example of how artificial intelligence is transforming familiar workflows and making powerful tools accessible to almost everyone.
P.S. In some countries, this app isn't all that necessary - people already know what's true.
Normally, journalists fact-check these statements, but their analyses often appear hours, or even days later, after millions of people have already watched and discussed the speech.
One student decided to change that by creating a Chrome extension called InTruth.
It runs directly on top of videos and fact-checks politicians' statements in real time.
First, the AI converts speech into text, then identifies claims that can actually be verified.
It searches for information from reliable sources, analyzes the findings using Claude, and displays the results almost instantly.
Instead of simply labeling a statement as "true" or "false,"
InTruth uses several categories: "True," "Mostly True," "Misleading," "False," and "Can't Be Verified."
Each verdict is accompanied by links to the sources, allowing users to review the evidence and draw their own conclusions.
What once required an entire team of journalists and fact-checkers can now be done by a single student using modern AI models.
It's another example of how artificial intelligence is transforming familiar workflows and making powerful tools accessible to almost everyone.
P.S. In some countries, this app isn't all that necessary - people already know what's true.