Lots of recent development in this area:
Image recognition (e.g. for commons and or suggesting images for articles):
http://ostatic.com/blog/google-open-sources-powerful-image-recognition-tool
Image Caption creation:
Detecting offensive content (maybe copyright):
http://www.digitaltrends.com/social-media/facebook-ai-screen-offensive/
Audio transcription:
http://www.theregister.co.uk/2016/10/19/microsoft_speech_recognition/
Pick yer poison...