Monday, May 20, 2002

Are you happy now?

After a week of political mudslinging regarding the gathering and dissemination of intelligence data, Dick Cheney this weekend went on national TV and said that more attacks against U.S. mainland targets are not only likely, but even possibly imminent. Well no, he didn't quite say that. What he said was there was a lot of "traffic" on the monitored communications channels, and intelligence from unnamed Taliban/al-Qaida detainees seemed to indicate another attack is at least in the planning stages.

Don't blame Cheney. His alarmist comments are only meant to cover the administration's asses in case something happens. So, when city 'X' is bombed, or a terrorist blows up a market in city 'y', the government can say "we told you this would happen". Gee thanks!

I don't know. Maybe it's that I'm Canadian, that my country's independence came through a stroke of the pen, that it's illegal to have handguns in this country, that until a recent friendly fire accident in Afghanistan, Canada had not lost a soldier in combat since Korea. Canadians are used to peace, and peaceful coexistence. So yeah, I get a bit nervous when people tell me that another mainland attack may be imminent. No offense, but I'd be much happier if the only nervous people around me were the government agents assigned to protect the country. They get paid to be nervous. I don't.

As if that weren't enough to screw up my day

To make matters worse, technology was shown this morning on the NBC Today Show that "literally" puts words into someone's mouth. Researchers at MIT have programmed a computer to morph the images on pre-recorded video tape so that subjects can be made to "appear" to say anything the author intends. The reported purpose of this technology is so that computer interfaces can be made more user-friendly, more human in appearance. I admit that the initial application of such technology will probably be focused on these purposes. After all, using the current state of software, the author of the morphing first has to capture sample "training exercise" speech from the intended human "target" on video. Just like phonemes, this library of video can then be cut into "snippets" representing each lip/facial combination, and these snippets are then combined with standard morphing techniques to "appear" fluid. As Katie Couric mused: those ads you see on TV, where long-dead superstars pitch modern day products, will now appear much more convincing.

At some point, these video snippets will become translated into generalized morphing "rules" (like video phonemes) that can be applied to the video images of "untrained" target images. How cool will it be to watch a foreign film, dubbed in English, where the lips on the actors actually match the soundtrack! But at that point, the technology can also be put to sinister use. People that have NOT volunteered their images for this experiment in deceit will be targeted. How long will it be then until terrorists are showing morphed images of heads of state - with closeups of their faces, as they spout uncharacteristically harsh words meant to inflame the masses? Or worse, when our own government shows incriminating videos of terrorists or other wanted suspects freely admitting crimes and discussing details that had never been uttered outside the cyber world.

To be sure, the simple manipulation of video is not enough. The current experiments are intended to put words into people's mouths - so the target subject's lips appear to utter the words heard on an audio track recorded by a completely different person. How long will it be until the audio track can be morphed to appear as if the words were spoken by the original person? How long until political leaders in hiding, dead or aging, are heard and seen to be extolling ideas and inciting actions long after nature would dictate this is not possible?

Now that is truly scary.

No comments: