i became convinced a while ago that broadcast news anchors were likely to be among the first victims of open, public-facing replacement by a combination of computer graphics and artificial intelligence. all the necessary technology to replace expensively spackled human newsreaders is pretty much ready for prime time — and if you are tempted to doubt this, or even to begin shouting “yay!” or “boo!” about a phenomenon as inevitable as the sunrise, check out this remarkable demo from a tech startup called channel1.ai . the entertainment news site deadline has a brief, only semi-illuminating primer on the company behind this showreel.
the whole thing has a character of envelope-pushing. the human reporters and techies at this shop are consciously exploring the question of just how much ai involvement in news broadcasts humans might ultimately tolerate. (be sure to keep ben sixsmith’s barbed joke in mind.) the demo presents an impressive (and racially diverse!) spectrum of artificial 3d human newsreaders: these avatars are not quite all the way out of the “uncanny valley,” as you’ll notice if you watch their hands, but they have some obvious advantages, like: they can start speaking any of about 200 languages at a microsecond’s notice; no bad hair days, ill-timed hangovers or personal scandals; and you don’t have to pay them.
advertisement
to me, this feels more like crossing a moral line than the avatar newsreaders do: subtitles strike me as a safer approach to ai translation in news broadcasts. then again, i remember my own past complaint that humans prayed for thousands of years for the blessing of universal natural-language translation , and have somehow completely forgotten about this dream at the moment it arrived. the argument whether ai image-and-sound “dubbing” is to be preferred over text subtitles … well, moviegoers have been quarrelling over that one for decades already, haven’t they?
advertisement
national post
twitter.com/colbycosh