Icky Fingers

When I was five or six my brother and I were very interested in the weekly Top Ten countdown on the radio. I can still remember our bafflement and disgust as week after week for almost two months Frank Ifield yodelling a song called "I Remember You" remained stuck at No. 1. What could it mean? Who were the people who could bear to listen to this appalling noise?

My memory has it that we wanted Ferry Cross the Mersey to knock off Frank Ifield - but it turns out it came out years later. If it wasn't that, perhaps it was The House of the Rising Sun. But, no: that too came years afterwards. So it must have been the Beatles with Love Me Do that we were barracking for.  I didn't think I liked that song much, but it seems that long ago a three-foot two version of myself cared about it very, very much.

In connection with our Top Ten mania, on Saturday mornings my brother and I used to go with our friends from nextdoor, Charlie and David, to the Chelsea Record Shop, on the Kings Road, almost next to the Chelsea Town Hall. Charlie and David had more pocket money than us (and television, and free access to sweets, such astonishing riches), and they would buy singles. I did not have enough money for singles - at least not genuine ones. However, there was a product that I could afford - a record that came out every week with the top ten songs on it, but all sung by people who, despite their best efforts, did not sound quite like the real thing.

There is nothing in my life that has been a bigger waste of money than those imitation hit records. I succumbed to the temptation of buying them three times, before realising they were just completely no good. Actually, I think I realised they were no good right from the first time I bought one - but the next couple of times I was genuinely interested to try to work out what it was that made them inferior. They were so very nearly like the originals. No one sang off tune. They had the lyrics and the tempo perfect, their singers had similar voices to those of the singers they were imitating. But they mysteriously failed - entirely - every time.

I remembered those records when someone showed me some of the photographs of non-existent people that artificial intelligence is beginning to produce. Here are a couple of examples:

They look remarkably good, at first glance. But look a little more closer - just like those fake top ten records, there is something not quite right about these fake people. It's their hands. What on earth is going on with their hands?

I find it reassuring actually. As someone on Twitter said:

Someone replied with this account of robots pitted against US Marines that was also quite cheering.

On the other hand, (ho ho), if Dora Carrington's portrait of Lytton Strachey is accurate, perhaps AI photograph generators have simply had Strachey-related data fed into their programming:

Only time will tell: will robots gain the upper hand (ho ho ho ho) or are their barriers we cannot get over? I grew up in an age when we were led to believe that humans could solve everything, but it's been 50 years and there still seems no solution to the problem of storing energy for any length of time in batteries. Either robots and artificial intelligence, as human-made phenomena, will reach certain limits and go no further, or they will go beyond those limits, not thanks to human efforts but surpassing them. I pray that the first option will prevail,  that robots and AI will reach the limits that we are capable of and go no further and that human life will meander on in its messy, muddled, mostly five-fingered way.

Ps: coincidentally (or because AI is already infinitely better than I'd imagined) straight after writing this, two articles on the subject of AI turned up in the ever flowing stream that is Twitter. Here is one. And here is the other. 

I Robot

While my husband was out for a couple of evenings last week, I watched Shoah. Possibly not something best done on one's own.

I haven't finished yet. I'm only five and a half hours in. But the things I've learnt so far aren't quite as straightforward as I'd imagined.

In fact, I haven't exactly learnt anything. Rather, I've found myself up against questions I can't answer about moral responsibility - and I'm also growing increasingly doubtful about Edmund Burke's assertion that "the only thing necessary for the triumph of evil is for good men to do nothing." I think Burke grossly underestimated the force that is bad men.

Not that it is always easy to define who exactly is bad.

Before watching the film, I would probably have said that all the many individuals who contributed in one way or another to the workings of Treblinka, Auschwitz et cetera were each unequivocally evil. After all, they had allowed evil to triumph by, essentially, doing nothing. Yet, if many of them - including, most troublingly, those inmates who were spared, provided they did appalling work within the system - had struggled against the powers in charge of the camps, what difference would it actually have made? Those individuals were not faced with the chance of saving someone else by sacrificing themselves. If they had refused to cooperate, all that would have happened is that they would have gone to their deaths and others would have been drafted in to replace them. Good men would have done something, been killed, and more good men would then have been drafted in - until some of those good men thought, "What's the point?"

I don't think cooperating allowed the triumph of evil, since not cooperating stopped nothing and resulted in extra death.

To begin with, though, I did imagine that it might be possible to differentiate between the moral responsibility of the person who wrote the detailed report that is read out at one point - a report specifying exactly what modifications were needed to increase the efficiency of the trucks that were used as killing chambers early on in the war - and the moral responsibility of the person who drove the engine that took the trucks into Treblinka. The report writer, it seemed to me, was in no real danger and was coldly, calmly applying their intelligence to how to improve the killing capabilities of machinery; the engine driver, by contrast, was in immediate danger -  should he refuse to drive his engine, he would be killed and someone else would be dragged in to do the driving instead.

His refusal would prevent nothing, whereas the report writer could have deliberately made recommendations that made the machines function in such a way that they didn't work at all. But actually, as I write, I begin to have my doubts about whether even the report writer had a real choice. Almost everyone had become trapped within a depraved structure of violence. It was the architects who created that structure - (those who, using violent intimidation, imposed violent methods on an entire nation) -  who were to blame. The report writer may have been one of these willing architects - or merely a frightened cog inside the death-dealing machine.

And so the only conclusion I can reach with certainty is that it is absolutely vital to ensure that you never ever allow your country's government to fall into the hands of people who consider violence a useful tool - or, worse still, actually get satisfaction or pleasure from violence.

But, as the lesson of history seems to be that that is virtually impossible, an alternative presents itself - one that has never until now been available. If humanity is not capable of ruling itself over a sustained period without resorting to violence - and this does seem to be clear from everything that has occurred in our history thus far - then perhaps what we need to do is put non-humans in charge. If we can manage to create a generation of robots that are programmed to have as their priority never ever allowing any kind of violence, then surely they will provide us with an ideal government. At a time when politicians seem to be held in lower esteem than ever before, could my gentle robots possibly make things worse?

But then I remember that other great phenomenon so unavoidable in human affairs. It is known as the law of unintended consequences. I withdraw all further proposals. While personally I think this 1750 French mannequin might make a fine President of the United States, I cannot guarantee it. I suppose it depends on what the alternatives turn out to be:

Why I Am Buying a Mask

A long train journey provides an opportunity to catch up on old New Yorkers. In a slightly too lengthy article on developments in artificial intelligence, in the issue of 19 January, I just read this passage::

"By scanning facial action units, computers can now outperform most people in distinguishing social smiles from those triggered by spontaneous joy, and in differentiating between faked pain and genuine pain. They can determine if a patient is depressed. Operating with unflagging attention, they can register expressions so fleeting that they are unknown even to the person making them. 

[Its co-inventor] often emphasizes that this technology can read only facial expressions, not minds, but Affdex is marketed as a tool that can make reliable inferences about people’s emotions—a tap into the unconscious."

Right - that is the last you'll see of me (inasmuch as you have seen anything of me to begin with) - I'm off to scour the Internet for a full-face mask.