Bad News for Democrats
Midterm elections are barely over and Craiyon already predicts the results for the 2024 presidential elections. The results won’t please everyone.
WHAT’S NEXT?
To the surprise of many bookmakers, America’s midterm elections have delivered rather good news for Democrats so far. Despite low approval ratings for President Biden, the Republicans seem to have suffered another defeat.
Since humans tend to make mistakes, it is time to ask an artificial intelligence and get rid of all doubts once and for all. Craiyon as a neutral spectator who draws images from any text prompt seems just born for this task. Dealing with different inputs concerning the upcoming election in 2024, Crayion comes up with a pretty strong opinion about the future president.
Even though the system has basic difficulties in face drawing due to limitations of its current image encoder, the lovely face of Donald Trump is quite recognizable and predominant. It remains interesting whether Craiyon already has the ability to look into the future or whether it is just a toy for adventurous geeks.
This is fascinating! I wonder what sources the AI was drawing on for this – why Trump vs Biden. Is there a skew in the sources the AI pulls from?
Wow, this is truly mindblowing. I wonder if the algorithm only picks up on parts of the prompt or overemphasizes the word election when it runs through its data and model to create the output… perhaps this could explain why all images clearly depict Trump.
I tried doing this and it’s unbelievable if you type “who’s going to win the election in 2024?” ( Hint : it shows the same person). In a way this if done correctly is a great tool for political campaigns. They’ll be able to understand the imagery out in the web pages and who’s leading on that end and what audience are seeing.. This can positively ( or negatively) direct their digital strategies maybe?
I think this is quite interesting. I wonder if asking for predictions from Craiyon sort of plays off a human flaw in human cognition. When we type in a prompt asking a question and images get generated (kind of like in our own minds), we think about the images generated as responding predictively versus simply scraping previous data histories related to the words in the prompt…