CAMBRIDGE, Massachusetts (AP) --
In an eerie video clip created by Massachusetts Institute of Technology researchers.

The MIT team has combined artificial intelligence and videography to make words and song -- even in foreign languages -- emerge from the lips of people who could never possibly have uttered them.

Potential for abuse.

Yet not only does the video blur past and present, it also heralds new possibilities for video mischief. Just as digital stills can be manipulated to misrepresent reality, so will advances in digital video technology enable full-motion fakery.

The new video sleight of hand is the work of a professor at MIT's McGovern Institute and Artificial Intelligence Lab.

Now, Poggio and Ezzat have programmed a computer to troll through short video clips and learn how a specific person speaks, a process that can take several days.

Once the computer has learned how the person shapes their mouth around individual sound segments it can digitally morph the shape of the subject's mouth around any audio sequence the creator wants to put words in a subject's mouth within minutes.

It's the "teaching" of the computer that makes this method different from most existing facial animation technology.

The recorded results? A woman made to sing in Japanese, and Marilyn Monroe lip-synching a song that didn't become famous until decades after her death.

Bob Steele, director of the ethics program at The Poynter Institute, a journalism research center, worries about the potential for abuse.

There are serious concerns that videotape could be doctored for unethical purposes: to fabricate evidence and literally put words in someone's mouth.

Consider, for example, all the tapes -- such as those from the Nixon presidency -- that have recently been released. How will the future authenticity of such audio recordings be guaranteed?

"If we use this new technology in a way that can alter reality, we certainly run the risk of deceiving the public, and if we deceive the public, we individually and collectively increase their skepticism of what they read in the paper and magazines, see on television or hear on the radio, or read online," said Steele.

Advances in digital imaging have already forced newspapers to adopt guidelines and protocols for altering photographs. Similar measures will need to be taken for facial animation, said Steele.

A study by Gadi Geiger, who authored the MIT research paper with Poggio and Ezzat, found that 22 adult viewers were only able to discern the real images from the animated images about half the time.

The MIT team does worry that its work, to be presented next month in San Antonio at Siggraph, might be used for mischief or worse. But its members say there's no more danger of misuse than with any other innovation, be it a computer program or a water pump.

Ezzat said video "watermarks" could prevent copying or alteration of tapes in the same way treasury bills are marked to prevent forgery.

Among those naturally concerned about video doctoring are lawyers.

James Alan Fox, a Northeastern University criminal justice professor, says video evidence will now need to be handled more carefully.

"It will not be insurmountable. With any new scientific wrinkle, there are hurdles that come about, but we resolve them, and I think that will happen here too," he said. "I still think that seeing will be believing."

I was going to post this a couple weeks ago, but I decided against it as this technology will be invaluable in the war against Eurasia.

You doubleplusungoodnik, we are at war with Eastasia.  We have always been at war with Eastasia.