I would like to propose what may be the most desperately needed modern application of Deep-Learning Artificial Intelligence: online fact-checking. If trained with access to a complete body of intelligence devoted to the truth, such as Wikipedia, machine learning could feasibly in real time respond to any and all user queries, and if possible, provide accurate and honest answers.
The current state of fact-checking capabilities are hindered by the high skill levels required, as well as by the shear volume of potential lines of inquiry. There quite simply are not enough fact-checkers with the education and ability to distinguish complex truths from fiction.
This is where Neural Networking would come into play. Regularly scraping internet archives would allow it to tap into a vast database of factual information, which it could then share with anyone curious enough to make use of its resources.
The truth, which has always struggled to be heard, would no longer be so vulnerable to disingenuity. And that would be a victory, not for any one person, but rather for humanity as a whole.
Add new comment