English | EN 
I have no doubt that Artificial Intelligence already walks among us. Figuratively speaking unless you worry about the impressive and yet still a bit weird robo-dogs you can view on streaming sites.
Last week, it suddenly appeared in a mundane, everyday situation. Not some eloquent chatbot that can tell me the weather, set a timer for me and tell me a lame joke. And not something akin to Hal 9000 or Skynet, both of which were recklessly programmed and clearly ignored Asimov’s three laws. So let that be a warning to you!
This was something simpler. The scales in my local supermarket have been upgraded. I use the term advisedly as the upgrade doesn’t fix the single most significant fault of the old scales. But more on that in a moment. These scales now have a camera that recognises what fruit or vegetables I have selected.
I overheard comments like “The machines are taking over” and “Where will it all end?” as I waited patiently for my turn. I found it mildly amusing that a set of scales which can differentiate between bananas and potatoes, caused such a response. The same people probably have devices at home, or in their pocket, with voice-recognition capabilities, and talk to chatbots when trying to solve issues with banks or utility suppliers, probably without giving it a second thought.
Now I assume the new scales are not really to help us mere humans but more as a revenue protection measure. As some less scrupulous customers may have been deliberately confusing lemons with watermelons to save a few pennies. However I’m fairly sure I probably still have the edge over weighing scales when it comes to image recognition. I can’t speak for all of my fellow shoppers, though. After all, some of them confuse lemons and watermelons… allegedly!
As an aside another supermarket also introduced another revenue protection measure whereby one must scan a valid receipt to escape. As long as you have a valid receipt with a bar code on it then you can escape. It doesn’t matter, as far as I could tell, if the items in your bags tally with the items on the receipt…
Week one was, for me at least, a raging success. No problems, fruit and vegetables correctly identified, and stickers provided for scanning. Week two had some issues, though. The thin mesh bags the supermarket promote instead of plastic bags for fruit and veg caused considerable confusion for the machine. And it was unclear how to resolve the issue from the confused look on the scale’s face (screen…). I did wonder if all these use cases had been thought of by the revenue protection team!
But there was more.
Week two and half the new machines had fallen to the same blight as their predecessors: Paper jams. It seems AI can be humbled, much like a decent cutting blade, by simple old paper!
Assuming one day the machines do overcome paper jams and other mundane IT issues we face, which machines will we trust, which machines will cause mild panic when introduced, and which machines should we be utterly terrified of?
Well, I suspect we’ll only know about the machines we ought to be utterly terrified of when it’s too late. I think it’s safe to say Hal 9000, Skynet, and the T-800 Terminator probably belong in that category, and there’s very little we can do about it.
Data from Star Trek was very like us and seemed to be well integrated into everyday life. As were, arguably, C3PO and R2D2 in their ways. It seems to be easier to get along with machines that follow Asimov’s three laws and which help us. And people like being able to ask the Internet questions via voice-recognition apps, so hopefully, they too abide by Asimov’s three laws.
So not only is it important to consider how we prove Artificial Intelligence is doing the right thing, but we also need to understand its motivation. Why was it created? And who does it serve? Which comes back to the big question of how do we test artificial intelligence successfully?
And as the change creeps slowly into are day-to-day lives, how do we manage the cultural change to ensure acceptance?
