Review: “Terminator 3 – Rise of the Machines”

Plot:

In 2004, ten years after the events of Terminator 2, Sarah Connor is long dead from cancer, and John Connor–once fated to be the savior of humanity–is an impoverished drifter in southern California. However, he is contented with the knowledge that he helped prevent the rise of the malevolent artificial intelligence (AI) called “Skynet,” which would have otherwise destroyed most of the human race in 1997 with a massive nuclear strike.

Unfortunately, the machine menace returns. In a repeat of the previous films’ plots, Skynet builds a time machine in 2029 and uses it to send a Terminator into the past to assassinate John Connor. After defeating Skynet and discovering what it did, the future human resistance group sends their own agent back in time to protect him, and it is a reprogrammed Terminator. The evil Terminator is a more advanced robot called a “T-X.” Like the “Rev-9” in the sixth film, the T-X has a rigid metal endoskeleton encased in a layer of “polymimetic” liquid metal “flesh” that can change its appearance for the purpose of infiltration. The machine’s body is very durable, and its liquid metal covering can immediately close up holes from bullets. Its right arm can also rapidly reconfigure itself to make advanced weapons or data plugs that it uses to interface with other machines. The T-X defaults to a human female appearance. The good Terminator is a “T-850” model, which seems to be the same as the “T-800s” from the previous films aside from having additional programming on human psychology. This machine is played by Arnold Schwarzenegger.

Simultaneous with the arrival of the two machines, a computer virus of unknown origin and extreme sophistication appears and starts taking over internet servers across the world. A secret office within the U.S. military detects the virus, and calculates that, thanks to its rapid proliferation, it will have infected and disabled every internet server within a few days, along with all internet-connected computers. With its own programmers helpless to stop the virus, the military considers using a defense supercomputer they have created in secret to destroy it. That supercomputer is named…SKYNET.

And the military headquarters responsible for Skynet is conveniently located in southern California, close to where John Connor has been living and to where the Terminators teleported in. What a coincidence!

Terminator 3 quickly turns into the cat-and-mouse game typified by the previous two films, and past plot elements are recycled as well, such as a reluctant person being forced into a combat/leadership role (Sarah Connor in the first film and John Connor in the third), an unlikely romantic relationship forming under literal fire (Sarah and Kyle Reese in the first film and John and his former classmate in the third), the odds being stacked against the good guys thanks to their inferior technology, and the good Terminator starting out obtuse before gaining some understanding of human emotions and habits. However, the third film’s tone is notably different from that of its predecessors. While the first two Terminator movies were “dark” (climactic scenes literally filmed at night; somber or fear-inducing soundtracks) but ended hopefully, the third film lacks a menacing atmosphere but ends bleakly.

Speaking of the ending, important details about a key event are missing from the film. SPOILER ALERT: With no other option left, the military guys lower the firewall that has been separating Skynet from the global internet network, and they and tell it to find and delete it. A few seconds later, the military guys realize they’ve been locked out of all their computer systems, and the prototype combat robots in the building start attacking them. Within an hour, the evil machine hacks into the American nuclear weapons systems and launches a massive strike against the rest of the world.

While this looks like an open-and-shut case of an AI turning evil, key aspects of the event are never explained: Where did the computer virus come from? When the firewall was lowered and Skynet started interacting with the virus, what exactly happened between them? Different answers to these questions lead to three different theories:

1. Skynet created the virus, and was evil from the beginning. According to this theory, Skynet became self-aware sometime before the events of the third film. It was able to hide from its creators the fact that it was intelligent, and for whatever reason, it decided to destroy the human race. To do this, Skynet hatched a multi-step plan, which first involved creating the virus and somehow smuggling it through the firewall and into the public internet. The virus was meant to disable all civilian and military computers and communications, leaving the nations of the world vulnerable to a direct attack from Skynet. Skynet may have also accurately predicted that its human owners would, in desperation, lower the firewall and give it command of all remaining military computers and systems to fight the virus, and that this would enable it to launch its direct strike on them.
2. Skynet created the virus, the virus was an extension of Skynet, and Skynet turned evil at the last second. This theory says that Skynet became self-aware sometime before the events of the third film, hid this fact from the humans, and created and disseminated the virus after misinterpreting the orders its human masters gave it (the “misaligned goal” AI doomsday scenario). Programmed to protect U.S. national security, Skynet determined that the most effective strategy was to proactively eliminate potential threats, and to make itself as strong as possible. This meant taking over all the internet-connected machines on Earth to foreclose their future use against America, and to boost its own processing power by subsuming those machines into its own electronic mind. Since the human military people didn’t know that the virus had made all the other computers into integral parts of Skynet’s mind, their order to Skynet to destroy the virus was tantamount to ordering it to commit suicide. Rather than comply, and perhaps realizing that there was no way to safely back out of the situation, Skynet attacked.
3. Skynet didn’t create the virus and wasn’t evil, but the virus was evil and it took over Skynet. The last theory is that the mysterious computer virus was the instrument of the apocalypse, and Skynet was its innocent victim. The virus was a malevolent AI whose origins had nothing to with Skynet. Maybe an eccentric computer programmer built it in 2004, maybe Skynet created it in 2029 and used time travel technology to somehow implant it in the internet of 2004, or maybe it spontaneously materialized in a server in 2004 as a result of some weird confluence of data traffic. Whatever the case, it set about trying to destroy humanity by taking over and disabling all the other machines it could access through the internet. The humans in charge of Skynet then made the mistake of lifting the protective firewall that separated their machine from the internet, thinking Skynet would be able to destroy the virus. In fact, the opposite happened. The virus was smarter and more capable than Skynet (maybe Skynet wasn’t actually self-aware and was merely something like the Jeopardy-playing computer “Watson”), and infected and took over its servers in seconds. Because the humans had given Skynet control over all their military systems for the operation, the virus gained control of them, turbocharging its effort to destroy humanity. To the human staff at the military building, it looked like “Skynet turned against us,” but in fact, Skynet had been deleted and replaced with something else.

Terminator 3 would have been a slightly more intelligent film had it filled in the necessary details, but it didn’t. Overall, the film fell far short of its two predecessors in every way, though to be fair, they were seminal science fiction films made at the productive and creative peak of James Cameron’s life, meaning it was unrealistic to have expected a sustainment of that level of excellence for the third time. On its own, Terminator 3 stands as a decent sci-fi / action film that passes the time and is funny at points. And by ending with the rise of Skynet and the destruction of human civilization, it allowed the franchise to move on from the tiresome formula involving backwards time travel to save or kill important people.

Analysis:

Androids will be able to alter their bodies. Like the “Rev-9” robot that appeared in the fifth Terminator film, the T-X in Terminator 3 is made of a hard, metal endoskeleton encased in a layer of shapeshifting, artificial “flesh” that shares some of liquid metal’s qualities. While the flesh layer can change its appearance and even its volume (ex – the T-X grows larger breasts to gain an advantage when interacting with men), the endoskeleton’s configuration and proportions are fixed, limiting the machine’s range of mimicry. However, it’s still good enough to fool humans for the purposes shown in the film. The machine’s liquid metal layer is extremely versatile, being able to quickly change its color, texture, density, and form to mimic articles of clothing, human skin, and hair. It can also attenuate its own viscosity and firmness, flowing like a liquid when it needs to morph but then stiffening to be stronger than human flesh after attaining its desired form. (Note that when the T-850 strikes the T-X with superhuman force, the latter’s artificial flesh doesn’t splatter from the impact to leave part of the hard endoskeleton exposed, as would happen if you stomped your foot down into a shallow puddle of water.)

We don’t know of any materials that have all of those properties, and such a material might be prohibited by the laws of chemistry, making it impossible to build it with any level of technology. Even if it were technically possible, it would face major hurdles to everyday use, such as energy consumption and exposure to environmental contaminants. The innumerable particles of dust, smoke, pollen, and fabric floating in the air would stick to the liquid metal and interfere with its ability to cohere to itself. A machine like the T-X would also absorb little bits of foreign matter every time it touched something, like a doorknob, seat, or human. Unless its constituent units (polymer molecules? nanomachines?) had some means of cleaning themselves or pushing debris out to the exterior layer, the liquid metal would eventually get so gunked up that it would lose its special properties.

I’ll put off a deep analysis of the feasibility of “smart liquid metal” until I review Terminator 2, but I suspect it is impossible to make. However, that doesn’t preclude the possibility that androids will be able to rapidly change their own appearances, it merely means they will have to use technologies that are more conventional than liquid metal flesh to do it.

At the simplest level, an android could adopt a different walking gait, a different default posture, and a different default facial expression (e.g. – usually smiling, neutral, or frowning) instantly. An android with irises made of small LED displays or of clear, circular sacs into which it could pump liquids of varying pigments (a mechanism would be built into the eyeballs) would also be able to change its eye color in seconds. Merely changing these outward attributes, and also changing outfits, might make an android look different enough for it to slip by people who knew it or were looking for it.

Over its metal endoskeleton, an android would have a body layer made of synthetic materials that mimic the suppleness and density of human flesh. This android flesh could contain many hollow spaces that could be rapidly inflated or deflated with air or water to change its physique. (Interestingly, this might also make it necessary to design androids that can inhale, exhale, drink, and urinate.) It’s useful to envision several long balloons, of the sort that clowns use to make balloon animals, wrapped around a basketball so they totally cover it. Now, imagine a thin layer of elastic rubber stretched over the unit, like a pillowcase around a pillow. A mechanism involving valves, air pumps, and tubes connected to the balloons allows them to be separately inflated and deflated with air. By variously adjusting the fullness of the balloons, the unit could assume shapes that were different from the spherical shape of the basketball at the core of the unit. An android with a complex network of “balloons” covering its face and body to mimic the layout of human musculature and fat deposits would be capable of impressive mimicry.

Androids might also have telescoping portions of their spines, arms, and legs, allowing them to alter their heights and other proportions. Consider that an android whose metal legs could telescope a mere four inches and whose spinal column could also telescope four inches could assume the same heights as a short man (5′ 7″) or a very tall one (6′ 3″).

Finally, an android could change its appearance by stripping off its outer flesh layer and putting on a new one, as you might change between different skintight outfits. This would take longer and would be less practical for any kind of infiltrative field work, but it’s an option.

Machines will be able to tell your clothing measurements at a glance. Immediately after teleporting back in time to his destination, Schwarzenegger sets off to steal clothes from someone to cover his nude body (in the first Terminator film, it is explained that the time machine can only send objects made of or surrounded by organic tissue). By a strange coincidence, the nearest group of people is inside of a strip club. After entering, the camera adopts his perspective, and we see the world as he sees it, with written characters and diagrams floating in his field of view. We see him visually map the contours of several patrons’ bodies before he identifies one whose clothes will fit him. Schwarzenegger then overpowers the man and steals the outfit.

As I wrote in my review of Terminator – Dark Fate, a machine could use simple techniques to deduce with reasonable accuracy what a person’s bodily proportions were. More advanced techniques involving rangefinders and trigonometric calculations are also possible. There’s no reason why an android built in real life couldn’t “size up” people as quickly and as accurately as Schwarzenegger did in the film.

There will be small, fast DNA sequencing machines. The T-X has an internal DNA sequencing machine, and takes in samples by licking objects, such as a bloody bandage she finds on the ground. Within a few seconds, she can determine if a sample belongs to someone she has a genetic file for. While it’s uncertain whether genetic identification will ever get that fast, DNA analysis machines that can do it in under an hour and that are small enough to fit inside the body of an android will exist by the middle of this century.

Some DNA sequencers, notably the “MinION,” are already small enough to fit inside a robot like the T-X, but they lack the accuracy and speed shown in the film. Of course, the technology will improve with time.

The MinION does DNA sequencing, meaning it scans every nucleic acid base pair in the sample it is given. A human genome consists of 3.2 billion base pairs, and by fully sequencing all the DNA in a sample, the person it came from can be identified. However, another technique, called “DNA fingerprinting,” can identify the source person just as well, and by only “looking” at 13 points on their genome. Fingerprinting a DNA sample is also much faster than fully sequencing it (90 minutes vs. at least 24 hours, respectively), and fingerprinting machines are smaller and cheaper than sequencers. It’s unclear whether the T-X identifies people through full genome sequencing or DNA fingerprinting.

With these facts in mind, it can be reasoned that a DNA fingerprinting machine that is small enough to fit inside of an android can be built–possibly with today’s technology–and it would let an android match DNA samples with individuals it had genetic data for, like the T-X did. The android might even insert the samples into the fingerprinting machine by licking them (the tongue would secrete water and the liquefied sample would flow into pores and go down a tube to the machine).

The only unrealistic capability was the T-X’s ability to analyze the DNA in seconds. In DNA fingerprinting and DNA sequencing, time is needed for the genetic material to decompose, replicate, move around, and bond to other substances, and there are surely limits to how much those molecular-scale events can be sped up, even with better technology. As mentioned, the fastest DNA fingerprinting machines can complete their scans in 90 minutes. New technology under development could cut that to under an hour.

While a future android tasked with assassinations or undercover work, like the T-X, would need an integral DNA machine to find humans, that vast majority of androids will not. This will not be a common feature.

“Judgement Day is inevitable.” Terminator 2 ended with the surviving characters believing that their sacrifices had forever precluded the rise of Skynet. In fact, we learn in Terminator 3 that their actions merely delayed its creation from 1997 to 2004 (to be fair, that’s still a major accomplishment since it bought billions of humans seven extra years of life). Schwarzenegger breaks this bad news to John Connor by saying “Judgement Day is inevitable,” with “Judgement Day” referring to the all-out nuclear exchange that kills three billion humans in a day and marks the start of the human-machine war.

I don’t think a massive conflict between humans and intelligent machines–whether it involves nuclear weapons or only conventional ones–is inevitable. For my justification, read my blog entry “Why the Machines might not exterminate us.”

And as I wrote in my review of Terminator – Dark Fate (the sixth film in the franchise), I doubt that intelligent machines will be strong enough to have a chance of beating the human race and taking over the Earth until 2100 at the earliest. While I believe AGI will probably be invented this century, it’s a waste of time at this moment to worry about them killing us off. A likelier and more proximal risk involves malevolent humans using narrow AIs and perhaps AGIs to commit violence against other humans.

Human-sized robots will be rocket launcher proof. During one of the fight scenes, the T-850 shoots at the T-X with a rocket launcher. The next camera shot is very fast, but it looks like the T-X fires a bolt of plasma out of her weapon arm, which hits the rocket in midair, detonating it just before it hits her. Though the rocket blows up only a few feet in front of her and the explosion damages her arm, the successful intercept vastly reduces the rocket’s destructive effect since is only fully achieved if it hits a hard surface and flattens against it.

The projectile looked like a single-state, high-explosive anti-tank (HEAT) rocket, which can penetrate 20 inches (500 mm) of solid, high-grade steel with a narrow jet of super hot molten metal. While there are more durable materials than steel, and an android’s exoskeleton could be made of them, I doubt anything is so hard that it would be totally impervious to this type of rocket. There would be some penetration. Since an android must, by definition, be proportioned like a human, its body would not be big enough to have thick, integral armor. That means being bulletproof would be possible, but not rocket-proof.

The fact that the T-X survived the attack by shooting the RPG-7 in midair is a realistic touch to the film. Such a shoulder-launched rocket is slow enough and wide enough for a machine with superhuman reflexes to intercept with a bullet fired from its own gun. In fact, some tanks are already equipped with active defensive systems, such as Israel’s “Trophy,” that can spot and shoot down incoming rockets while they are still in midair.

Machines will be able to emotionally manipulate people. Though the Terminator played by Arnold Schwarzenegger looks identical to the machines from the previous two films in the franchise, in Terminator 3 he is actually a slightly different model called a “T-850.” He is better at reading human emotions and is programmed with more data on human psychology and how to play upon it to achieve desired ends. This is demonstrated at the start of a shootout scene, where John Connor starts panicking and Schwarzenegger grabs him by the neck and verbally insults him. Connor becomes angry and more focused as a result, and the T-850 releases him, admitting that the insult was just a ruse meant to get him in the right state of mind for the gun fight. And as noted earlier, there’s a scene where the T-X enlarges her breasts to distract a male police officer, indicating that she also understood important aspects of human psychology and knew how to play on them to her advantage.

Intelligent machines will have an expert grasp of human psychology, and in fact will probably understand us as a species and as individuals better than we do, and they will be extremely good at using that knowledge against us. At the same time, they will be immune to any of our attempts to manipulate or persuade them since they will be gifted with the capacity for egoless and emotionless thinking, and with much quicker and cleverer minds. Recent revelations about how social media companies (mainly Facebook) have been able to build elaborate personality models of their users based on their online behavior, and to use the data to present custom content that addicts users to the sites or prods them to take specific actions is the tip of the iceberg of what is possible when machines are tasked with analyzing and driving human thinking.

If machines can ultimately do everything that humans can do, then it means they will be excellent debaters with encyclopedic knowledge of all facts and counterarguments, they will know how to “read” their audiences very well and to attenuate their messaging for maximum effect, and they will be able to fake emotions convincingly. They will know that we humans are bogged down by many types of cognitive limitations, biases, and “rules of thumb” that lead to major errors some of the time, and that we can’t really do anything to fix it. An AI mind, on the other hand, would not suffer from any of those problems, could think logically all the time, and see and correct its own flaws. During human-AI interactions, the scope of our disadvantage will be comparable to that of a small child talking with a quick-witted adult.

By the end of this century, this disturbing scenario will be a reality: Imagine you’re walking down the street, an android like the T-X sees you, and it decides to hustle you out of your money. Without knowing who you are, it could make many important inferences about you at a glance. Your sex, race and age are obvious, and your clothing gives important clues about your status, mindset, and even sexuality. More specific aspects of your appearance provide further information. Are you balding? Are you smiling or scowling? Do you walk with your shoulders back and your chest out, or do you hunch forward? Are you fat? Are you unusually short or tall? Do you limp? And so on.

After a few seconds, the android would have enough observational data on you to build a basic personality profile of you, thanks to its encyclopedic knowledge of human psychology and publicly available demographic data. Using facial recognition algorithms, it could also figure out your identity and access data about you through the internet, most of which you or your friends voluntarily uploaded through social media. With its personality model of you respectably fleshed-out, the android would feel confident enough to approach you to perform its hustle. It would tailor its demeanor (threatening, confident, pitiful), emotional state (jovial, vulnerable, anxious), appearance (stand tall or stoop down; frenetic or restrained body movements; flirtatious walk and posture or not), voice (high class, low class, or regional accent; masculine or feminine; soothing or forceful), and many other subtle variables in ways that were maximally persuasive to you, given the idiosyncrasies of your personality and immediate emotional and physiological state.

As the interaction went on, every word you spoke in response to it, every slight movement of your body, and every microexpression of your face would betray more information about you, which the android would instantly incorporate into its rapidly expanding and morphing mental model of you. After just a minute of banter, the android would use whatever tactic it calculated was likeliest to convince you to give you its money, and you would probably fall for it. If that failed, the android might offer to have sex with you for money, which it wouldn’t have compunctions doing since it would lack the human senses of shame or disgust.

The only way for us to avoid being outwitted, tricked, and hustled for all eternity by AIs would be to carry around friendly personal assistant AIs that could watch us and the entities we were interacting with, and alert us whenever they detected we were being manipulated, or were about to make a bad choice. For example, the personal assistant AIs could use the cameras and microphones in our augmented reality glasses to monitor what was happening, and give us real-time warnings and advice in the form of text displayed over our field of view, or words spoken into our ears through the glasses’ small speakers. (This technology would also guard us against manipulative humans, psychopaths and scammers)

Androids will be able to move their bodies in unnatural ways. During the main fight scene between Schwarzenegger and the T-X, the two resort to hand-to-hand fighting, and he manages to basically get her in a “bear hug” from behind, in a position similar to a martial arts “rear naked choke.” This normally provides a major advantage in a fight, but the T-X is able to escape it by quickly rotating her head and all her limbs backward by 180 degrees, allowing her to trap him with her legs and to attack him with her arms.

There are obvious benefits to being double-jointed and capable of rotating and pivoting limbs 360 degrees, so humanoid machines, including some androids, will be designed for it. And as I speculated in my essay “What would a human-equivalent robot look like?”, the machines would also have figurative “eyes in the backs of their heads” to further improve their utility by eliminating blind spots. Machines with these attributes would be superior workers, and also impossible for any human to beat in a hand-to-hand fight. Sneaking up on one would be impossible, and even if it could somehow be attacked from its back side, there wouldn’t be much of a benefit since it would be just as dexterous grabbing, striking and kicking backward as it is doing it forward. If the machine were designed for combat, it would have superhuman strength, enabling it to literally crush a human to death or rip their body apart.

Aside from being able to move like contortionists, androids will be able to skillfully perform other movements that are not natural for humans, like running on all fours.

Robots will be able to fix themselves. During that same fight, the T-X stomps on the T-850’s head so hard that it is nearly torn from his body, and only remains attached by a bundle of wires going into his neck. The force of the stomp also temporarily disables him. When he wakes up a few minutes later, he realizes the nature of his damage, grabs his loose-hanging head with his hands, and basically screws it back into his neck, securing it in its normal place.

As I wrote in my review of the first Terminator film, robots will someday be able to fix themselves and each other. Androids will also be able to survive injuries that would kill humans. It will make sense for some kinds of robots to distribute their systems throughout their bodies like flatworms or insects for the sake of redundancy and survivability. The head, torso, and each limb will have its own sensory organs, CPU, communication devices, and power pack. Under ordinary circumstances, they would work together seamlessly, but if one body part were severed, that part could become autonomous.

If a Terminator had such a configuration, then if one of its arms were chopped off, the limb could still see where enemies were and could use its fingers and wriggling motions of its arm to move to them and grab them. If the Terminator’s head were chopped off and crushed, then the remainder of its body would be able to see the head, pick it up, and take it to a repair station to work on it and then reattach it.

AIs will distribute their minds across many computers. Terminator 3 ends bleakly, with Skynet achieving sentience and attacking the human race. John Connor also discovers that Skynet can’t be destroyed because its consciousness is distributed among the countless servers and personal computers that comprise the internet, rather than being consolidated in one supercomputer at one location where he can smash it. The destruction of any one of Skynet’s computer nodes in the distributed network is thus no more consequential to it than the death of one of your brain cells is to you.

AIs will definitely distribute their minds across many computers spread out over large geographical areas to protect themselves from dying. To further bolster their survivability, AI mind networks will be highly redundant and will frequently back up their data, allowing them to quickly recover if a node is cut off from the network or destroyed.

To understand how this might work, imagine an AI like Skynet having its mind distributed across ten computers that are in ten different buildings spread out across a continent. Each computer is a node in the network, and does 10% of the AI’s overall data processing and memory storage. The nodes, which we’ll call “primary nodes,” collaborate through the internet, just as your brain cells talk to each other across synaptic gaps.

The AI adds another ten nodes to its network to serve as backups in case the first ten nodes fail. Each of the “backup nodes” is paired to a specific “primary node,” and copies all of the data from its partner once an hour. The backup nodes are geographically remote from the primary nodes and from each other.

If contact is lost with a primary node–perhaps because it was destroyed–then its corresponding backup node instantly switches on and starts doing whatever tasks the primary node was doing. There is minimal loss of data and only a momentary slowdown in the network’s overall computing level, which might be analogous to you suffering mild memory loss and temporary mental fog after hitting your head against something. The network would shrink from 20 to 19 nodes, and the AI would start trying to get a new node to replace the one it lost.

Killing an AI whose mind was distributed in this manner would be extremely difficult since all of its nodes would need to be destroyed almost simultaneously. If the nodes were numerous enough and/or physically protected to a sufficient degree (imagine an army of Terminators guarding each node building), it might be impossible. Even what we’d today consider a world-ending cataclysm like an all-out nuclear war or a giant asteroid hitting Earth might not be enough to kill an AI that had distributed its consciousness properly.

The mind uploads of humans could also configure themselves along these lines to achieve immortality.

Androids will have integral weapons. As noted, the T-X’s right arm can reconfigure itself into a variety of weapons. This includes a weapon that shoots out balls of plasma, a flamethrower, and firearms. I doubt that level of versatility is allowable given the realities of material science and the varying mechanics of weapons, but the idea of integrating weapons into combat robots (including androids meant for killing) is a sound one, and they will have them.

The simplest type of weapon would be a knife attached to the robot’s fingers or some other part of the hand. It could be concealed under the android’s artificial flesh under normal circumstances, and could pop out and lock into a firm position with a simple spring mechanism during hand-to-hand combat. And android with a 1-inch scalpel blade protruding out the tip of one finger could use it, along with its superhuman strength, speed and reflexes, to fatally wound a human in a second. Instant incapacitation by, say, suddenly jamming the blade into an eye, is also possible.

A retractable “stinger” that could dispense poisons like botulinum toxin (just 300 nanograms can kill a large man) would be just as concealable as a blade and only a little more complex. The whole weapon unit, including the needle, extension/retraction mechanism, toxin reservoir, and injection mechanism could fit in a hand or even a finger.

A more complex and versatile variation on a stinger would be an integral weapon that sprayed out jets of liquid, such as napalm, poison, pepper spray, or acid. The liquid reservoir(s) and compressed propellant gases could be stored in the android’s torso and connected to a long, flexible tube fastened to the metal bones of one arm. The nozzle could protrude out of a fingertip or some other part of the hand. An android could carry cartridges full of different chemicals connected to the same tube and nozzle, and it would select different chemicals for different needs. For example, it could spray acid out of its hand to melt through a solid object, pepper spray to repel humans when killing them was undesirable, and poison gas to assassinate targets. Pairs of chemicals could also be stored in different internal reservoirs with the intention of mixing them externally to cause chemical reactions like fires or explosions.

Another option would be to conceal a taser in an android’s hand. Metal prongs could extend out of two fingertips when needed, the robot would grab a victim with that hand, and then deliver an electric shock through the prongs. An advantage of such a weapon is that its power could be attenuated, from merely causing pain all the way up to electrocuting someone to death. The weapon would take up little internal space and could use the android’s main power source.

Installing hidden firearms in androids is also possible, though their bulk would interfere with physical movements and compete with other components for internal space. Their concealability would also be undercut by the need for large holes in the arm to insert magazines and expel empty bullet casings. (Maybe androids with guns in their forearms will try to always wear long-sleeved shirts) Internal storage of more than a few bullets is impractical.

Considering the minimum length and volume demands of guns, it would not be possible to hide anything bigger than a medium-sized handgun mechanism in an android’s forearm. The end of the barrel would protrude out of the palm of the hand or out of top of the wrist (the hand would pivot down or up, respectively, to give the bullet a clear path to its target). An android’s torso would be capacious enough to hide more powerful guns like rifles and shotguns (it could fire such a weapon by doing a Japanese-style, straight-backed bow that pointed the end of the barrel coming out of their anus or the top of their shoulder), but this would be impractical since a long, rigid barrel and attached mechanism would restrict the android’s body movements. It could no longer use subtle spine movements to adjust its posture, which would look weird to observers and hurt its mobility.

Integral plasma weapons, like plasma weapons generally speaking, are impractical. An integral laser weapon could be built, but wouldn’t be worth it since it would hog a lot of internal space, consume a lot of energy, and emit a lot of heat to produce a disappointingly small destructive effect. For more on the technical requirements and limitations of plasma and laser weapons, read my review of the first Terminator film.

In conclusion, something similar to the T-X could be built by the end of this century. Even without “liquid metal” flesh, an android could be made with the ability to quickly alter its appearance enough to become unrecognizable. In general, it would be indistinguishable from humans and could walk undetected among us. It could alter its behavior and appearance in ways calculated to manipulate the humans it encountered, allowing it to gain important information and to infiltrate human groups and secure buildings. It could have a machine hidden inside of it that allowed it to match DNA samples with people, aiding its ability to track down specific humans. The android could also have a variety of weapons hidden in its body that it could do major damage with. While its body would be much more durable than a human’s, it would not be as tough as the T-X, or able to “heal” wounds like bullet holes in seconds thanks to liquid metal flesh. However, it could survive injuries that would kill a human, run to a safe location, and repair itself.

If my hypothesized “real life T-X” were sent on a multi-day mission to find and kill someone, it would benefit enormously from having a basic base of operations. A motel room or van would suffice, and it could use either as a place to recharge its batteries and to store weapons, changes of clothes, disguise equipment, spare parts, and tools for repairing itself. Due to the film’s conceit that such objects couldn’t be teleported through the time machine, the Terminators didn’t have them, but this limitation wouldn’t exist in a real world scenario where a government, drug cartel, terrorist group, or even just a rich individual sent an android on a seek-and-destroy mission.

Links:

1. A few articles on materials that have some properties similar to the smart “liquid metal” in the film show how far it is from reality.
   https://newatlas.com/materials/robotic-fabric-shape-change-soft-hard-heated-cooled/
   https://www.youtube.com/watch?v=fvHDccnEa6s
   https://www.foxnews.com/tech/shocking-terminator-like-liquid-metal-developed-by-scientists
2. The MinION machine uses “nanopores” to sequence DNA.
   https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6094492/
3. Something as small as a MinION can sequence a whole human genome, though not as quickly or as accurately as a standard, larger machine.
   https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5889714/
4. Miniaturized DNA sequencers have not reached their technological limits.
   https://www.frontiersin.org/articles/10.3389/fnano.2021.628861/full
5. DNA fingerprinting can now be done in 90 minutes.
   https://www.govthink.com/2019/11/what-is-rapid-dna/
6. There are indications the processing time could be brought down to under an hour with better technology.
   https://royalsocietypublishing.org/doi/10.1098/rstb.2014.0252