This will be resolved very subjectively by my impression. I will try to mitigate this subjectivity by resolving to multiple responses in cases where I am unsure.

This question will primarily depend on "perceived nationwide rate of behavioral/lifestyle changes resulting from this application". Relevant behavioral changes would include, non-exhaustively:

1. spending time talking with a romantic companion AI

2. changing news consumption patterns as a result of widespread AI

3. additional free time available to commuters using fully autonomous self-driving vehicles

4. reduced engagement with human sources when defaulting to question-answering AI over wikipedia, Quora, or primary sources

5. increased travel if speech-to-speech AI make travel more appealing (unlikely to be significant enough at the everyday level)

6. Major catastrophe caused by AI

7. Technical AI (e.g. coding, doctor etc.) becomes usable to untrained users.

8. ...

I will also take into account metrics such as "number of headlines on this application", "number of friends who engage with this application", "willingness of users to pay for this application", "magnitude of effect in terms of minutes/day engagement with this application", google search statistics, etc.

I will roughly take a "difference-in-differences" approach to measuring this impact. To clarify what I mean by this consider news: news has a large impact on peoples' opinions. However if all newspaper writers were replaced by AIs, this may not have a large felt impact, since the difference-made-by-news could remain the same. In such a case, I would not resolve this question to "News AI" since the difference-in-differences of this application would be minimal.