This is not doable.

Unless hacked, your smartphone doesn't communicate with other smartphones around it. It doesn't even know if someone is taking pictures. There's no way it could possibly block someone taking a photo of you.

In the clip you've linked, A explains that he can do "whatever he wants", like "altering the Internet connection". He's implicitly saying that he hacked her phone, that's how he got full control of it.

Here is a patent for technology that will tell smart phone A not to take a picture of asset B.

https://www.theatlantic.com/technology/archive/2016/07/what-if-cameras-stopped-telling-the-truth/491150/

Now if all the smart phone vendors implemented this, similar to how the consumer grade DVD player vendors all agree on DVD copying then eventually you'd get the situation where most phones don't take the image. Not perfect of course.

I think there might be scenarios in which this could work. Imagine a person wearing a T-shirt with a QR code. Person B tries to shoot a photo with his smartphone. The smartphone recognises the QR code which

1. 
   

   automatically opens a web browser showing a webpage stating "Stop
   this"

   
   
   

   or

   
   



2. generates a buffer overflow due to buggy implementation of the
   QR code interpreter.

If A has knowledge of an unpatched vulnerability in B's phone, then it's possible that A's device could wirelessly upload an exploit to B's phone and interfere with the photo software's operation to prevent the photo being saved. If this is done passively, then A needs exploits for any phone that they could encounter. It also wouldn't work on older digital cameras that don't have wireless capability, but most people don't carry those anymore. The ability could only be available to a select few, since software and/or hardware manufacturers would patch the vulnerability if they got their hands on the exploit.

It's possible than an exploit like this exist, but it'd either be in the hands of a government agency using it for surveillance or someone who would sell it.

From the video clip, it doesn't look like there is physical interference with the camera, since you can see the target's face on the screen before the photo is taken. MechMK1 gave some good examples of physical interference methods, but there is also the case of a billionaire's yacht blocking paparazzi cameras by shooting lasers at the sensors and this purse that reacts to camera flashes with its own

Part of this is theoretically possible, but the scenario as described is a TV fantasy.

There is active research into detecting lenses pointed at a target, but they are aimed at sniper rifles, not cameras, e.g. this patent. There are also multiple ways to detect hidden cameras, aimed at surveillance. And, of course, there's even an app for that.

These technologies could be combined into a detector that notices when someone points a camera at you, including that from a smartphone. Leaving questions of reliability and false positives aside, it could be combined with a technology like the anti-face-detection LEDs MechMK1 mentions to form a kind of selective camera camouflage.

But how would you get the other person's smartphone to show a message, or a picture entirely different from the one they were trying to take? The first case would require you to hack their smartphone, over the air, in real time. If that were feasible, it would already happen, and for more nefarious purposes. For a spy movie, that's the kind of rocket-shooting, underwater-driving car that we accept in such movies, but it's not real.
To show a completely different picture (the 2nd example in the clip) simply defies the laws of physics. You would have to project light into the camera aimed at you, but to fill the entire picture and not the tiny dot of your projector lens would require the light beams to bend in ways that light simply doesn't bend like.

Yes, it is doable, but it would be a voluntary feature of the camera app of the person that takes the photo. I do not think it is implemented in any common camera app, but there are implementations in many photocopiers, that prevent you from copying things with a certain dot pattern on it.

The EURion constellation is a special pattern, that can be recognized by these devices and stops people from copying banknotes. In addition to photocopiers, some software like Adobe Photoshop implements this as well.

People already experimented with printing EURion on t-shirts and other clothing. I do not think they were successful.

Other experiments involve adversarial pattern for image recognition neural networks, that prevent face detection or lead the camera not to focus on the face, when it is in auto-focus mode. Most of such patterns are rather obvious make-up, that you would not wear every day. Another technical device that uses infrared LEDs is mentioned in the post of MechMK1.

This is not something which anyone can attain. Previous responses explain very well why.
However there are some people which may actually go very very near to that!
Most modern smartphones either work using iOS (Apple) or Android (Google).
If you were the CEO of Apple you cloud simply impose that the company forces a silent update to all devices (phones/tablets) so that if they detect your specific Bluetooth they will refise to take pictures!
A similar thing can be obtained from the Google CEO but it would require a more tricky approach (no need for detsils: your engineers will know how to do that :-)
So to sum it up to block all the phones you need the cohoperation of both Google's and Apple's CEO, then a simple app on your phone will simply trigger the protection!
:-)

So....
In case (very unlikely!) these company already have embedded such "no photo" trigger and you are the first and only hacker to discover it then you could simply implement the client app to trigger this!

If the original question is limited to "hacking" just Person B's phone, then it's quite doable with only minor "suspension of disbelief."

First, Person A obtains B-phone's SIM id. Then Person A's phone checks the location of PhoneB, and if it's within, say, 10 meters, PhoneA checks an outgoing message (here's where some hack software would have to have been loaded) from PhoneB indicating a camera is activated. Now Person A is at least warned.

FWIW, it is completely reasonable to posit that NSA-level tools can force OTA installs onto PersonB's phone without any notifications being made. And for that matter, given another few years' worth of software development and smartphone processor capability, why not claim that such a software install includes facial recognition tools that could identify PersonA and then intercept the current camera image(s).