The exception for news media in the EU strikes me as interesting. I think it will be quite tough to verify adherence to this without a firmer idea of what qualifies as proper "editorial review" or oversight. Thanks for laying this out so clearly.
This was an insightful piece, Nick. As a strong advocate for transparency, I generally support AI labeling and giving consumers the information they need to decide what they engage with.
I also suspect we’ll see a growing premium placed on distinctly human-created work. Just as handcrafted products command higher value today, content created with minimal AI assistance may eventually become a differentiator in its own right.
It will be fascinating to watch how policy evolves—particularly around the New York Act and whether regulation ultimately aligns with what consumers actually want: transparency, accountability, or some combination of both.
All policies pertaining to deepfakes are naturally quite challenging. Anecdotally, editorial guidelines for publishing deepfake-related incidents require very stilted incident titles. Example:
Incident 544: **Alleged** Use of **Purportedly** AI-Generated and Manipulated Media to Misrepresent Candidates and Disrupt Turkey's 2023 Presidential Election
"Alleged" because we can't often attribute intent and we have not ourselves verified the facts through independent reporting. "Purported" because we can't say for certain that something is a deepfake or that someone has not taken a real video and inserted deepfake elements to produce doubt.
The exception for news media in the EU strikes me as interesting. I think it will be quite tough to verify adherence to this without a firmer idea of what qualifies as proper "editorial review" or oversight. Thanks for laying this out so clearly.
This was an insightful piece, Nick. As a strong advocate for transparency, I generally support AI labeling and giving consumers the information they need to decide what they engage with.
I also suspect we’ll see a growing premium placed on distinctly human-created work. Just as handcrafted products command higher value today, content created with minimal AI assistance may eventually become a differentiator in its own right.
It will be fascinating to watch how policy evolves—particularly around the New York Act and whether regulation ultimately aligns with what consumers actually want: transparency, accountability, or some combination of both.
All policies pertaining to deepfakes are naturally quite challenging. Anecdotally, editorial guidelines for publishing deepfake-related incidents require very stilted incident titles. Example:
Incident 544: **Alleged** Use of **Purportedly** AI-Generated and Manipulated Media to Misrepresent Candidates and Disrupt Turkey's 2023 Presidential Election
"Alleged" because we can't often attribute intent and we have not ourselves verified the facts through independent reporting. "Purported" because we can't say for certain that something is a deepfake or that someone has not taken a real video and inserted deepfake elements to produce doubt.
https://incidentdatabase.ai/cite/544/