While we often report facts and throw in our viewpoints from partners, the next story is a bit surprising. Last week China reported on the “world`s first” artificial intelligence news anchor, a story that quickly went viral online. This presenter at Chinese state Xinhua News Agency was praised about “his” ability to report on broadcasts 24 hours a day while never getting tired nor needing a break. Now after being reviewed, Will Knight, a senior editor at MIT Technology Review is skeptical.
The likeness is, in the words of experts, a real human, somehow being adjusted for purpose. A sub-field of A.I. or a new kind of machine learning seems to be added to the mix. He said that the image does not have any intelligence, and the lips moving and other images are using altorithms that are related to A.I. but in fact are not. Calling this an A.I. anchor is slightly overstating it.
Some A.I. is able to look through mountains of documents and give views on how to address the content. For example, Imperial College in the UK, has A.I. review and analyze over 100,000 medical records and work out, for doctors to view, the best way to give patients specific treatment.
The ability for a machine to view and understand some human patterns is here, but to read a newspaper then write a review or give an opinion is surely difficult at this point. For the most part, machines can still make mistakes. As these technologies are introduced and become more human-like, we will be reporting with of course, our viewpoints here on our platform.
Life-like images do look like humans, but behind images is not a thinking robot but and algo putting together memorized comments and allowing for facial movements – so the anchor was not thinking, just repeating what he was told – likely by a human.