iPhone X is out. The most hyped feature on this mean machine is its 3-D facial recognition feature which authenticates the user before the phone can be opened and used. In other words, you can now say good bye to fingerprint, PIN or password based authentication. 

Apparently, the feature is pretty accurate and it cannot be fooled by disguises or using masks. It captures some 30,000 points of your face and creates a 3 D map and stores it in a chip within the device in encrypted form for security.

Facial recognition technology is getting better and better, though there are potentially huge security and privacy issues involved (and which we shall discuss in a later article). But for now, let’s do some blue sky thinking on at least one beneficial way facial recognition can be used in tandem with other technologies and artificial intelligence.

To catch a thief or to prevent a murder

Well I do not really mean to be so dramatic, but can we ignore the realities that we see around us today?

I had an uncle who was retired and rich and would spend a good amount of time sitting in the balcony reading newspapers and magazines. But any suspicious goings on in the street below would not escape his eye. His vigilance stood him and the neighbourhood in good stead.

But such human dependent vigilance can be flawed and is likely to miss important events once in a while. Enter the heroes of the new era - CCTV camera and Aadhar – the last being considered more a villain than a hero nowadays.

Imagine that someone is recceing my house with the intention of theft or perhaps murder. (No harm in imagining things!!) The person comes on his bike and runs down the road glancing at my property, turns around and goes back, glancing again all the way. He does this a few times over the day or over a few days.

There are several CCTV cameras on the road and outside my house as well. All of these are connected to a data centre where an AI algorithm runs on the videos being fed to it in real time. The AI algorithm has ‘learnt’ over the past several months or years to know that this person’s frequent movement in front of my house could be a ‘suspicious’ event. Very much like what my uncle would have surmised.

Now, that is what AI is meant for. It is supposed to learn by experience what we as human beings have learnt by experience and come to same or similar conclusions on its own and even faster than what we can do.

The system then raises an alarm since it has detected an anomaly – in this case the alarm can come to me, my neighbors and even the police - via phone call or SMS or a buzzer or whatever.

Now imagine that the police have access to the data center server where the video gets streamed and have this integrated with the Aadhar database. Now, Mr. Thief, if he has to live in this country, must definitely have his Aadhar stored there – right? In all likelihood, if the CCTV footage is reasonably good, another piece of AI algorithm will be able to find in minutes, if not seconds, the right match between the face on the CCTV footage and the Aadhar photograph.

I leave the rest of this whodunit to be completed by you, dear reader.

Can the technology go wrong? Of course, it can. Perhaps the chap was actually innocent. But then my uncle’s hunches could be wrong as well. But that does not mean that uncle would not at least get up and verify and assure himself that all was OK. Same is with the AI system.

Well, all of this is now in the realm of possibility though not used in the fashion as described in this story. Such intelligent technologies can ensure better safety and security to citizens at a cost/effectiveness ratio that is much lower than current methods. But there could be privacy and legal issues to be grappled with to make this happen.