All the security money can buy

Just read this article via Twitter and it gave me chills for all the wrong reasons. The article was originally published in Forbes.

I haven’t fully processed this in my head but I wanted to get down why I thought this was a terrible way of looking at safety and security in the urban environment. (Going to assume you’ve read the article first.)

I’m open to having my mind changed on any of the following, by the way.

Joshi sets the scene talking about perils to society (“…organized terrorism, accidents, planned robberies or thefts…”), putting the cause of these down to “other people”.

Here’s my first problem, there’s a marked lack of curiosity about how these perils originate and what motivates the behaviours. The piece is written in a way that encourages the thought that these terrible things are here, we don’t need to think about why they’re here, we just need to do something to prevent them taking place. Oh, look! Technology!

Security as a commodity?

And preventing such incidents is paramount among the duties of smart city governors, along with ensuring the adequate supply of basic necessities such as food and water. To that end, smart city governments can use technologies like IoT and artificial intelligence in physical security.

Naveen Joshi (2019)

And there’s my second problem; treating security and safety as a commodity; something that needs to be supplied to the population (or the bit of it that isn’t doing the bad things). And like water or food supply, it’s something that is delivered by technical infrastructure, in this case AI and Internet of Things.

Then rest of the article is basically a breathless run down of the surveillance state’s greatest hits.

Security as an indicator of society’s health?

But what if we made the assumption that security isn’t a commodity to be supplied, but the by-product of a well-functioning and just society?

I’m no criminologist but I think we left the idea that “some people are just bad” behind us a long time ago. Reasons for criminality and radicalism are hugely complex with roots in economic hardship, mental and physical health problems and social disenfrachisement.

Using technology to cure the symptom won’t address those causes

But that’s OK. They’re “other people”.

Also, by turning safety into a commodity rather than a right or responsibility, it becomes subject to market forces and inequality. Ask the people of Flint, Michigan what it’s like to be denied a basic commodity.

Looking at the list of tech solutions Joshi promotes to solve a safety and security crisis, it’s not unforeseeable that, mishandled, they could exacerbate problems of urban inequality, lack of social and economic well-being and so on.

And another thing!

There’s a race and class side to this that I’m not qualified to talk about but I can’t help noticing that comparatively few people are talking about using AI and machine learning to identify white collar crime before it happens or to flag up political and corporate corruption.

Or maybe there’s a blockchain for that.

Photo by Matthew Henry on Unsplash

There’s something very odd about this movie trailer

Watch this trailer. It’s for the upcoming film Morgan and it’s very interesting. How effective do you think it is? Does it feel like all the other trailers you watch or different somehow. HINT: Don’t rewind it as the preamble gives some spoilerific context.

You probably picked up on the clues before you got the explanation at the end of the trailer (the title overlay gives it away).

The trailer was pieced together mainly using artificial intelligence. IBM’s Watson computer analysed the full movie and constructed the 1 minute plus trailer from what it thought were the key moments. There was some human creative input but it’s a bit unclear as to what that was.

Read the full post from The Next Web.

Beyond the eerie nature of the film, I found it a bit unsettling. I watch a lot of trailers and without knowing all the ins and outs of how they’re constructed I feel like I’ve internalised what a Hollywood style trailer should be like. I suspect most of us have – try watching one from a movie over 15 years old and you’ll probably be surprised how different they are.

This one, while seeming on the surface to follow the right patterns, doesn’t feel quite right. It’s a similar feeling to looking at one of those “realistic” androids where the features are all in the right place and the skin looks properly textured but as soon as it starts to move we’re in the uncanny valley.

What really interests me is how well AI can understand narrative. While a trailer only gives a partial story it still tries to convey a sense of the meaning of the story so you come to the cinema already understanding what the film will be about. How people do that is highly subjective and I don’t think easily computable. It’s a bit like the Turing test in that sense (and maybe a computer won’t be able to pass the Turing test until it understands narrative properly).

We’re also feeding the computer a lot of rules about what we think a good trailer should look like and Watson is emulating that. It’s not the same as creatively coming up with an original and creative way of doing the same job (although you could argue that’s exactly what some human marketing people in Hollywood are doing anyway!)

In this case, I don’t think Watson gets it quite right. It’s like a computer working through a checklist of things that work in theory but still manages to miss some important beats. Some cuts happen too early or too late and the pacing feels a bit wrong but for all sorts of intangible reasons I can’t fathom, annoyingly.

But…

…it still works.There are still some brilliant moments. The scene where Morgan talks to Paul Giamatti about his daughter, for instance. That genuinely made my skin crawl with the editing meaning the camera relentlessly focussed on his discomfort. Also the moment when the chair doesn’t quite make it to smash the window before the cut to black creates real tension. A human would be proud of those.

So all in all, the unnatural, unsettling feel suits this movie right down to the ground.

It would be interesting now to see how it handles a rom-com or a costume drama.