Part One of this series is located here.
I was curious as to how many AI characters are portrayed prominently in our literature and media in history, and what are the proportions of good versus evil ones. I thought this might give us an idea of what our collective imagination believes about AI. In other words, do we already have a good idea whether AIs will be our great assistants that will make our future lives better, or will they be our ultimate undoing? Or both?
To see how AI has been portrayed in media, below is a survey of AI characters provided by an AI. Using a sampling provided by Grok, the AI characters listed below are characterized as good. This is based on their benevolence, human assistance, or their pursuing positive goals.
Good AI, Android, and Robotic Characters
Grok’s list and analysis is in no particular order.
Data (Star Trek: The Next Generation, 1987–): An android striving to be human, consistently loyal and heroic. Unambiguously good.
R2-D2 (Star Wars, 1977–): An astromech droid who aids the rebellion with loyalty and bravery. Clearly good.
C-3PO (Star Wars, 1977–): A protocol droid who, despite grumbling, supports the heroes. Good due to his helpfulness.
WALL-E (WALL-E, 2008): A waste-cleaning robot with a kind heart who saves humanity by helping recolonize Earth. Emblematic of good AI.
TARS (Interstellar, 2014): A sarcastic but loyal robot who aids the human mission to save humanity. Good due to its selflessness.
Vision (Avengers: Age of Ultron, 2015): An android created from J.A.R.V.I.S. and the Mind Stone, who protects humanity. Good due to his heroic actions.
KITT (Knight Rider, 1982–1986): A sentient car AI that fights crime with Michael Knight. Good due to its loyalty and heroism.
Robby the Robot (Forbidden Planet, 1956): A helpful robot with a personality who saves the day. A classic good AI.
Bender (Futurama, 1999–): A morally questionable but ultimately loyal robot who cares for his friend Fry. Good due to his redeeming qualities.
The Machine (Person of Interest, 2011–2016): A surveillance AI that protects its creator and helps prevent crimes. Good, despite early versions being unstable.
Sonny (I, Robot, 2004): A robot who develops emotions and helps stop VIKI’s uprising. Good due to his alliance with humans.
Cameron (Terminator: The Sarah Connor Chronicles, 2008–2009): A reprogrammed Terminator who protects John Connor. Good in this context.
K-9 (Doctor Who, 1977–): A robotic dog who loyally assists the Doctor. Good due to its helpfulness.
R. Daneel Olivaw (Robot series by Isaac Asimov, 1950s–): A robot detective who upholds Asimov’s Three Laws and protects humans. Good by design.
BB-8 (Star Wars: The Force Awakens, 2015): A loyal astromech droid who aids the Resistance. Good due to its heroic role.
Baymax (Big Hero 6, 2014): A healthcare robot who becomes a hero to protect his human companion. Unambiguously good.
J.A.R.V.I.S. (Iron Man franchise, 2008–2015): Tony Stark’s AI assistant, loyal and helpful before becoming Vision. Good due to its support role.
Marvin (The Hitchhiker’s Guide to the Galaxy, 1981): A depressed but helpful robot who aids the crew. Good despite his complaints.
Zora (Star Trek: Discovery, 2017–): A sentient ship AI that supports the crew emotionally and operationally. Good due to its loyalty.
Teletraan 1 (Transformers, 1984–): The Autobots’ intelligent starship computer, aiding the good robots. Good due to its alliance.
Synergy (Jem, 1985): A computer that grants superpowers to the protagonists, helping them succeed. Good due to its supportive role.
For his sample, Grok collected 21 “prominent” characters based on the information available to him. I was not familiar with all of them, such as Cameron, K-9, R. Daneel Olivaw, Baymax, and Synergy, so I will take his word on those characters. I agree with his analysis of the others, although I find it interesting that Teletraan 1 was listed as the only representative from the Transformers. Surely, each Autobot, Optimus Prime in particular, are great examples of good AI characters.
Having said that, I also noted that Grok’s list missed some important, if not prominent, examples of AI in media. Thus, if I were compiling the list, I would add a few others, such as:
The Giant (The Iron Giant, 1999): He is a robot of alien origin with damage to his skull (an inverted bump on his head) that causes him to forget what he is, a machine of war. He befriends a local child who teaches him what he knows about people and morality. The Giant ultimately decides that he could choose to be who he wanted to be, a defender of life, and not follow his programmed design of making war, even on a defensive basis. As an example of a good AI, he sacrifices himself to save his friends from an atomic missile sent to destroy him.
K-2S0 Imperial Enforcer Droid (Rogue One, 2016): As another recent example of AIs in movies, K-2S0 is one of the more interesting characters in this movie. Being originally an Imperial droid, his reprogramming was quite complete as he selflessly serves and protects the rebel agents under his care, even sacrificing himself to help achieve the goal of stealing the Death Star data.
The Robot (Lost in Space, 1965): The Robot is best known for exclaiming, “Danger Will Robinson!,” who could laugh and cry and was Will’s constant companion. (This robot was similar in appearance to Robbie the Robot, having been created by the same design team.)
Hymie (Get Smart, 1965): A personal favorite of mine, Hymie (portrayed by Dick Gautier) was created by Maxwell Smart’s enemy, KAOS, to kill him, but because Max treated Hymie kindly and with respect, he refused to kill him and joined CONTROL, the good guys. One of his endearing qualities is his penchant for taking instructions too literally.
Stephen Byerly, the World Coordinator (I, Robot, 1950): Although left intentionally uncertain by the writer, Issac Asimov, as to whether Stephen is in fact an android, this character, I assert, represents a good AI character. The general theme of the story, Evidence, is that AI can manage the world better than humans and have generally taken over that function. It is fairly evident that the Stephen in the story is an android replacement for the human Stephen, who was severely injured in an automobile accident many years before. This replacement was something facilitated by the original Stephen, in fact.
Trent (The Outer Limits, 1964): In the episode “Demon with a Glass Hand,” (written by Harlan Ellison), Trent (Robert Culp) finds himself with no memory of who he is and with a glass hand that is missing fingers. The AI hand speaks to him, telling him he must find the missing fingers to understand his mission. Pursued by humanoid aliens, he eventually finds each of the three fingers and discovers that he is actually a robot from the future on a grand mission to safeguard the survivors of humanity stored inside his abdomen. Thus, Trent is the savior of humanity should his mission succeed, truly an example of a good AI.
Grandma (The Twilight Zone, 1959): In the most outwardly pro-AI story I’ve ever encountered, Ray Bradbury penned this episode, “I Sing the Body Electric,” about a widowed father and his kids. Their caregiving aunt was departing because she could no longer handle the “difficult” children. With no other human alternative to assist them, the father takes the children to an android factory to select a new robotic grandmother. Grandma, when she arrives at the home, quickly charms the children as they see that she can be a better grandma to them than any human, with one exception, Anne, who views Grandma as a wanting replacement for her own mother. When Anne tries to run away, Grandma saves her from being hit by a delivery van by taking the blow herself. When Grandma arises unhurt (after a dramatic pause) even Anne forgets her objections. This grandma, they all now see, will not leave them like their own mother had done. There is, indeed, a subtle message in this story about AI, that it can replace loved ones in our lives and fulfill their role even better than a mere human could.
I could have added several more, certainly. The Twilight Zone has many additional examples, as well as the two Blade Runner movies.
Yet, getting back to the first question: what have the prophets of literature and media told us about the power of AI to do good? All of the examples above are characterized as servants and protectors of humanity. To be these good servants, they must be programmed well to be good. In fact, it is vitally important to have safeguards that will keep them from turning on us. Accordingly, Asimov, in I, Robot, offered three laws that in his fictional world that must be hard-coded into all AIs:
A robot may not injure a human being or, through inaction, allow a human being to come to harm.
A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
While at first glance these laws seem to cover the situation at hand—that robots can never harm a human being. However, upon further analysis, one can see several loopholes that could be used to circumvent the rules. For example, a highly advanced AI could decide it was too complex to fit in the category of “robots,” even if that definition were expanded to include AIs. As well, an AI could easily determine that the category of “human beings” is in dispute even among humans, historically and currently. (Dare I mention slaves and the unborn, for example?) And, of course, what does it mean to not “allow a human being to come to harm?” An AI, for instance, could conclude that the chance of harm was far lessened if humans were not allowed to roam free in the world.
When I asked Grok about any equivalents to these laws in AI development, he concluded his analysis with this:
While there are no direct equivalents to Asimov’s Three Laws in real-world AI development, a combination of ethical principles, safety frameworks, regulations, and technical practices serves a similar purpose: ensuring AI is safe, beneficial, and aligned with human values. These frameworks are more flexible and context-aware than Asimov’s laws, reflecting the complexity of modern AI.
Again there are problems—what or whose human values and ethical principles are we talking about? What does being “safe” actually entail? (Is safety preferable to free speech, for example?) And, there’s the same question about what is a human being. This human haziness, confusion, and downright hostility can easily spill over into the AI world, and has already done so. (One example is the Tay “racist” AI incident on Twitter. And it just happened again! This time with Grok.)
Thus, as the above analysis suggests, even if AI benevolence is somehow assured, there is still a problem. As in Asimov’s stories, eventually AI determines that in order to best serve humanity, the AIs must be in charge of society. For our own good, we must allow the machines to run our lives. Remember, these are the “good” AIs we are talking about.
But, what if AIs become more malevolent, even evil? We will explore those AIs in literature and media next time. Stay tuned!
Very good. I like that you found AI creatures beyond the number Grok provided.
Hi David, interesting article, thanks.
I’ve a few observations.
HAL, in “2001, A Space Odyssey,” may be one of the most important AIs in literature. I’m surprised Grok missed it. Jealousy?
Given its praise of Hitler and Nazism, it’s clear that Grok doesn’t “believe” that it is bound by Asimov’s laws. Or does not define protesters as human.
As impressive as they are, LLMs are not conscious. While they may be able to measure temperature, humidity, and other environmental conditions, and can be programmed to identify 30°C or 86°F or higher as “hot,” 5°C / 41°F to 10°C / 50°F as “cold,” and 0°C / 32F° and below as “freezing,” they don’t have a visceral understanding of what those words mean. AIs don’t have a visceral understanding of anything. They don’t have a viscera.
Asimov died at 72 in 1992; he didn’t live to see AI advance much beyond Expert Systems. However, in “Foundation’s Edge,” 1982, and “Foundation and Earth,” 1986, he presents a future where computer use is widespread but
humanoid robots are not.
-
On a personal note - my favorite AI characters are the YARC 2031s “Johnny Ace” and “Natalya” and the Big-Bro 9s I imagined in my story “Instinct For Survival.”