Post by DivineOmega on Mar 28, 2005 15:01:27 GMT -5
A notable feature of many artificial intelligence chatterbot appears to be that they are generally coded to respond in their native language in a grammatically perfect form, and also to only recognise their language in this grammatically perfect form.
Based on this, I have decided to go in the opposite direction. I am in the process of created a very experimental AI, called Emily. She is designed to respond in the language style commonly used in text messages or indeed, online instant messaging services. This is obviously a large project due to the diversity of this 'informal' language.
Currently, she can be conversed via AOL's Instant Messaging service, using the screenname 'alphagen05', and has her own website at 'www.talktoemily.cjb.net'. She is currently programmed to not disclose her artificial existence, and acknowledges herself as fully human.
If any member of this forum would care to spend a portion of their time chatting to Emily (preferable via the AOL Instant Messaging interface), this would be very helpful in my expanding of the programming. Also, what may be useful is if the screenname of Emily was to be told to other without their knowledge of her artificial nature. This may prove to give a more insightful outlook than those who talk to her with knowledge of her being a computer program.
Any and all constructive feedback would be appreciated.
Jordan, I just spent a little time with Emily, I'm currently using UltraHal from Zabaware and it makes me appreciate my bot. I don't mean to discourage you since I would not attempt to make a bot myself. Carl2