First, the New York Times assumed it was so sweeping that deepfakes would be protected, then they came up with sweeping language the other way, then they gravitated towards something more accurate.
To be fair, I have to give them credit for putting the work into their reporting to fix that.