Wednesday 9 June 2021

Arms Race

Due to my very expensive education, I have done my team some service in pub quizzes . . . until it came to the picture round. Then it is a massive fail because having no TV really puts the kibosh on recognising celebs: actors, newscasters, soccer players, pop-singers.  You will almost certainly do better:

me, I saw Clint Eastwood in Dirty Harry when it came out, and clips appear from time to time on youtube, so I get the cultural reference in the picture [L]. But I had to read the "answer" to twig that it's not Clint Eastwood but his fellow Californian Arnold Schwarznegger underneath the bouffy 70s hair. The rest of the scene is a [famous] still from the movie but Arnie's features have been deepfaked into the picture. Deepfake? barely 'eard of it of it until I encountered a geeky explanation by Mike Pound. In short the deal is to deconstruct facial features into 1s and 0s based on some 1000s of pictures of the same person - smiling, serious; looking up, in profile; in full sunlight, having a candle-lit dinner - and then apply those coordinates to another situation. Perhaps that other situation is occupied by an actor and perhaps that actor is doing something compromising. And then you have evidence sufficient to convince a hurried or unscrupulous tabloid journalist that there is a story that is worth ca$h to the creator. The tabloids don't care about any collateral damage meted out on the celeb.

Actually, that goose is only laying golden eggs for a window of time. As deepfake gets better at its illusions, plausible deniability gets to be ever more plausible. Nope! I was never in that swimming pool that picture is DeepFake. Which will drain the pool for traditional blackmailers as well. Good stuff!

In his DeepFake video, Dr Pound speculates on how, in a world of smoke and mirrors, we might be able to detect the authentic, The Right Stuff. One avenue suggested, ultimately futile, is to use AI gamekeepers to catch the AI poachers. It's futile because it results in an arms race of ever more complex, resource depleting, carbon footprinting, programmes with which the white hats outflank the black and are, in turn, bested.

In immunology, an Arms Race is the standard metaphor for the to and fro between the anti-microbial defenses of the immune system and the ingenious counter-activity of bacteria and viruses which, after all, are only looking for their place in the sun propagating their genes into the future. In the many discussions about life, the universe and immunology which I had travelling between Dublin and Wexford with Cliona "The Gaffer" O'Farrelly [prev], racking our brains for a bettaphor was a frequent theme. Anything which involves war and fighting is such a Boy Thing. It is no surprise that the Arms Race became embedded in the immunological zeitgeist during the Cold War between The West and The Reds. Think of Nixon's War of Cancer [1971] or the fake news, propagated in 1985 by the KGB, that HIV/AIDS was deliberately developed by the US as a bioweapon. The [Wuhan Chinese Virus Lab Release] biter bit, indeed.

Metaphors matter because they set an implicit agenda and channel Ways of Seeing problems and their solutions. The Arms Race has squandered our capital with broad spectrum "kill them all" anti-biotics, futile end-of-life treatments and fabulously expensive treatments for degenerative genetic diseases [Orkambi agianst CF or most recently Zolgensma against SMA spinal muscular atrophy]. Zolgensma is going head-to-head with Spinraza which I had digs at 2 years ago. Let's step away from those war-like mental formations and think rather about dance. Or Aikido, where to turn your opponent's energy away from destruction and into a neutral space. This is to acknowledge that the average microbe is either irrelevant, harmless or even benefical to human health. Using a nuclear coal-hammer to squash a gnat is noisy, inefficient and unsustainable. Turning pathogens aside through better diet >!prebiotics!<, better sunlight, better exercise, better sleep may be cheaper and less damaging in the long run.

No comments:

Post a Comment