Is it possible to train our LINK AI to also identify possible risky themes or creative elements?
Consumers today tend to be very sensitive about different things such as gender, race, religion, moral and safety issues, politics these days so having a mechanism in AI that flags those will be helpful.
One idea would be to feed our AI with recent news or topics of discussions, general sentiments, that can help it flag if a certain element, theme or narrative may trigger certain kind of responses from viewers?
Additional details:
In a recent discussion with Samsung, they are hoping to have a Risk Sensing feature in the LINK AI as a way to integrate their Social Buzz/Listening data along with a Risk Sensing metric alerted by LINK AI.
One idea is for the AI to be trained using online comments, news and other source of social trends and context to generate a metric e.g. 'Risk Factor' that can be used as a signal if there is potential negative themes or issues used in the ad.
This can be an initial way about it but an even more sophisticated approach as further development could be to factor these online buzz/news/positive and negative sentiments data into how the ad performs.
Is there anything similar or adjacent to this that is being conceptualized or considered for development?
Guest
@Guest , Just remembered that I also logged this idea before. I updated the title to also note that it's for Samsung.