How do the programs know that someone is not plugged into the Matrix?

Question

Throughout the Matrix movies we are continually shown that an Agent could be anyone anywhere.

In Neo's training we are shown that anyone could possibly become an Agent, by example of the women in the red dress.

If that were the case, wouldn't anyone at any time change into an Agent, meaning that they could not mingle with anyone that was still plugged into the Matrix?

Then there is this scene where Neo, being someone who is still plugged in, does not change into an Agent, although they pull a gun on him to make sure that if he did they would be able to protect themselves.

That is unless there is a way the Agents detect someone who is not plugged in and that method was not very accurate.

Which got me thinking: why does the garbage truck which is seemingly just passing by know that Trinity is not plugged in and therefore turn into an Agent?

So, is there any explanation on how the Agents pick up that someone was not plugged in? And why is it so inconsistent?

Answer 1

Agents can only possess a person plugged into the Matrix (a "bluepill"):

[Agents] can move in and out of any software still hard wired to their system. That means that anyone we haven't unplugged is potentially an Agent.

_{Morpheus, The Matrix}

Zion operatives, however, hack into the Matrix via a wireless signal that they broadcast from their hoverships in the real world. We know this is a wireless signal because the hoverships are seen fleeing sentinels while Zion operatives are jacked into the Matrix (e.g. near the end of The Matrix, when the Nebuchadnezzar is nearly destroyed by sentinels before Neo is able to jack out of the Matrix).

Therefore, the Agents can determine if someone is plugged into the Matrix by whether or not they can possess him. More simply, the Agents can determine if someone is plugged into the Matrix by looking for a wired connection to that person. In the specific case where Trinity escaped just before the garbage truck (driven by Agent Smith) hit her: Agent Smith was able to determine that there was no wired connection to Trinity and that he could not possess her. Consequently, he possessed someone nearby who was hard wired to the Matrix (the bluepill driver of the garbage truck) instead.

---

[W]ouldn't anyone at any time change into an Agent, meaning that they could not mingle with anyone that was still plugged into the Matrix?

To answer this, I will borrow from my answer to a question about the purpose of the bug used by the Agents:

Agents can "see" what any bluepill sees but since there are billions of bluepills it is impossible to look in real time at what all of them are seeing to pinpoint the location of Zion operatives anywhere in the Matrix. This exhaustive search can only be narrowed down if a bluepill sees something unusual (e.g. an operative breaking the rules of the Matrix) and/or if Agents know the rough location of the operatives (such as when they are in a chase, in which case they can check only what is seen by the relatively few bluepills in that area). In the case of the vagrant in the subway, the Agents had been chasing Neo and Trinity so they knew roughly where they were; they undoubtedly checked what the vagrant saw (especially since he was in view of a phone the operatives could use to jack out) and Agent Smith possessed him when he recognized Neo and Trinity.

It's not that the Agents' method of determining whether or not someone is plugged in is inaccurate: it's that there are far too many connections (bluepills) to monitor them all in real time. The Agents require some means of narrowing down the number of connections to check (an anomaly, like a supernatural ability exhibited by a Zion operative). Zion operatives can safely "mingle" with bluepills in the Matrix so long as they don't do anything unusual which draws the attention of the system.

Answer 2

From what I can tell, programs are alerted to abnormalities in the Matrix.

Ie, When Morpheus disappears through the phone in the subway scene, the homeless guy turns into an agent due to the abnormality,

Any of the other times the people turn into agents are also when abnormalities happen in the Matrix

Answer 3

I don't think at anytime throughout the entire trilogy there is any indication that an agent obtains hard data about unplugged users. If someone is unplugged, then they simply do not exist in the matrix. So by "not plugged", I think you meant to say someone who is "non-genuinely authenticated" -- or a red user. So now we ask, how does a program know if someone is a red user?

We know agents can override blue user sessions remotely. As far as agents overriding red user sessions remotely, I believe they probably could do this if the red user's location was known; and the red user is not protected by a firewall. In the case with Bane and agent Smith, Smith is shown to be capable of overriding a red user. This was done non-remotely which may have aided in the breaching of his firewall.

With that being said, I agree with @KyloRen's response in that agents use metadata based on abnormalities correlated with geolocation and crowdsourcing metrics. Upon an abnormality, a particular agent is instantiated to investigate the locale by overriding the particular blue user pinged closest to the geolocation of said abnormality. At this point the agent isn't necessarily looking for a red user but simply investigating the abnormality. Beyond this point, based on what can be seen in the films, an agent can visually identify a red user as a non-genuine user session; this is evident. Perhaps the machines have integrated automated monitoring of ongoing non-genuine user session logins but the data would be meaningless to the agents without geolocation.