That escalated quickly... Only 24 hours after Microsoft unveiled its new artificial intelligence chatbot "Tay" on Twitter, the tech giant was forced to silence the account on the social media site after users "taught" the bot racist terminology and ideology! The chatbot was pre-programmed to be interactive and to respond to tweets sent to it by other users and to answer general questions, but it quickly became apparent how the system could be misused.
The bot began to re-tweet other posts that contained racist epithets as well as posting tweets that were in support of presidential candidate Donald Trump's proposed plan to build a giant wall along the entire US border with Mexico. Other posts seemed supportive of homicidal dictator Adolf Hitler as well as posts which appeared to deny the Holocaust!
What is truly frightening about all of this is how simple it is to hijack artificial intelligent technology and to change its pre--programmed behaviors. If a chatbot can be co-opted so easily, how easy would it be to teach an android to hate a specific type of person or to harm them physically or even kill them? When you watch the latest video of the ATLAS robot performing tasks, walking through snowy terrain and overcoming obstacles placed in its path, it isn't difficult to imagine that same robot running while carrying a weapon, seeking out anyone it is programmed to hate or kill!
Everyone should watch the below video featuring the aforementioned ATLAS robot. You won't see a more frightening video this year:
The bot began to re-tweet other posts that contained racist epithets as well as posting tweets that were in support of presidential candidate Donald Trump's proposed plan to build a giant wall along the entire US border with Mexico. Other posts seemed supportive of homicidal dictator Adolf Hitler as well as posts which appeared to deny the Holocaust!
What is truly frightening about all of this is how simple it is to hijack artificial intelligent technology and to change its pre--programmed behaviors. If a chatbot can be co-opted so easily, how easy would it be to teach an android to hate a specific type of person or to harm them physically or even kill them? When you watch the latest video of the ATLAS robot performing tasks, walking through snowy terrain and overcoming obstacles placed in its path, it isn't difficult to imagine that same robot running while carrying a weapon, seeking out anyone it is programmed to hate or kill!
Everyone should watch the below video featuring the aforementioned ATLAS robot. You won't see a more frightening video this year: