38th Carnegie Symposium on Cognition Paper

Summary By Grok:

Robert Thibadeau’s talk presents Perceptron Theory as the fundamental model of neural computation in both human brains and modern humanoid AI.

A perceptron is the basic computational unit: weighted inputs are summed at a moment in time to produce an output that feeds other units (illustrated by the “hand metaphor” of fingers as inputs and the thumb as output, following the labelled-lines principle).

This is the simplest correlation/recognition function.

The human nervous system contains hundreds of billions to quadrillions of such units (countable via methods such as fMRI).

Over 1,000 neuron types all perform the same basic function so they can interact. Weights constitute neural memories — either learned (e.g., cortical) or inherited (e.g., retinal).

These operations appear directly in natural language as subject-predicate pairings, predications, and contextualizations, a view now supported by AI models.

Thibadeau traces his career of building computational models of humanoid brains under the Turing Test to explain neural computation:

1973 multimodal memory work (verbal + pictorial) showing distributed memory across neurons.

1976 PhD on causal inferences between sentences, using probe reaction times as a direct window on cognitive computation and treating natural language as a window on neural processes.

1977–79 action-perception studies (with belief-system and expectation-based models) that emphasized sensory levels and used Lisp-based perceptron systems.

1979–82 eye-movement studies of skilled reading with Marcel Just at CMU, leading to CAPS (Collaborative Activation-based Production Systems) as a perceptron-style model; this revealed the need for massive specialized parallel hardware.

Later work as Lab Director at the Robotics Institute reinforced multi-layer perceptron pattern recognition (the specific correlation function is secondary to Turing completeness).

A railroad-car recognition system demonstrated near-perfect large-scale performance and illustrated the brain’s massive parallel overkill; this echoed his great-grandfather’s eidetic railroad-car memory.

At Seagate he embedded safe Lisp interpreters and perceptron linear algebra in disk drives for natural-language search and storage, anticipating that trillions of memory points enable humanoid AI.

Subsequent roles involved self-encrypting drives, computer security, and humanoid SaaS systems; fMRI work with Just further confirmed that neural distance matters because functional computations must be spatially close for survival-level speed.

Modern neural-network AI (LLMs, generative systems) consists of layered perceptron models. A fundamental symmetry exists in causal/agency computation: two perceptrons are required to “tango.”

Computational linguistics (e.g., inter-sentence causal inference realized in transformers) contributed to this Turing success.

Since 2018 Thibadeau has elaborated the theory through 95+ papers and videos on TruthCourt.net, introduced Mendaciology (the computational cognitive science of lying and deceit), and called for AI laws that assign personal responsibility to builders and users.

TruthCourt.net is presented as a SaaS platform for structured trials on truth/lie claims in media, using linguistically universal verdict questions, free for educational use.

The talk closes by crediting Marcel Just with keeping the path open for this lifetime body of work in computational cognitive science.

The powerpoint slides are attached. Note grok only had the original pptx file and read and summarized it in 2 seconds.

Original Powerpoint Slides

Here is the text of the newspaper article on JR Turner the Eidetiker mentioned above and on Slide 8, provided by Grok and confirmed as accurate: https://rhtcmu.medium.com/jr-turner-eidetiker-6eaddb830278

Perceptron Theory was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

Willful Ignorance Should Be One Word In English

Truthcourt.net is how to stop Amathia! Please sponsor trials so the Amathiasts can learn for a few bucks to stop themselves from being willfully ignorant!

This guy talks about Amathia creating evil dictatorships

https://www.facebook.com/share/v/1Pfyei79gW/

AI: Amathia (ἀμαθία) is the Greek philosophical word that means willful ignorance — not simple lack of knowledge, but a refusal to learn, often by someone who could know better and is sure they already do. Plato describes it as ignorance plus the arrogant belief that one is already wise. Socrates treated it as a serious moral failing, worse than ordinary ignorance.

Amathia is a special kind of lie. It is a lie of omission by refusing dialogue or other ways to learn about the truth as other people understand it.

If we adopt “amathia” as an English word, as in “he has amathia,” we can have some interesting derivatives:

• “amathic” as in “he’s being amathic” or “he’s amathic”
• “amathist” as in “he’s an amathist” believer
• “amathism” as in “he practices amathism”
• “amathology” the study of amathism or willful ignorance

On the Internet YOU HAVE TO PAY YOUR SOCIAL MEDIA PEOPLE TO LEARN!!! Please help us pay everyone on earth a small amount to learn how not to be stupid.

Sponsor some trials at truthcourt.net and anonymously meet your rewards.

Amathia was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

This is a Lie

I meant the title to be “Thumper” not “Trumper.” That’s the lie.

I’ve gotten really upset with YouTubes that lie in their titles in order to get you to watch them.

But now that I’ve already published “Trumper,” I fear that Trump might “read” this and wave my picture and title while he’s walking to his helicopter telling the news media about his new supporters.

But what kind of lie is this? Is this a socially acceptable lie? This particular kind of lie is worth investigating because of what is going on in your brain right now!

To Grok: tell me twenty ways in which “I’ve decided to become a trumper.” can be lie

“Here are twenty distinct ways in which the statement “I’ve decided to become a Trumper” can be a lie:

1. The speaker has made no decision at all about political affiliation.
2. The speaker has actually decided not to support Trump or identify as a Trumper.
3. The speaker was already a long-time Trump supporter, so there is nothing new to “become.”
4. The speaker is only pretending for social, professional, or relational advantage and does not believe it.
5. The choice was forced by external pressure (family, job, peer group) and was not a free personal decision.
6. The speaker does not know or understand what “Trumper” means and is using the word incorrectly.
7. The statement is pure sarcasm or irony meant to convey the opposite.
8. The speaker has decided to vote for Trump once but explicitly rejects the label “Trumper.”
9. The speaker plans to act like a Trumper publicly while privately holding the opposite view.
10. Someone else made the decision for the speaker (e.g., a campaign handler, spouse, or handler).
11. The speaker is an AI, bot, or scripted account and therefore cannot “decide” anything.
12. The speaker has decided to support a different candidate or party entirely.
13. The decision was made in the distant past and is no longer operative; the present-tense claim is false.
14. The speaker intends only a temporary or situational pose, not a genuine conversion.
15. The speaker is testing reactions or trolling and has zero actual commitment.
16. The speaker is role-playing or performing for an audience and does not mean it.
17. The speaker confuses “Trumper” with an unrelated meaning (card player, someone who “trumps,” etc.) and is not referring to politics.
18. The speaker has decided to stop being a Trumper but is saying the reverse.
19. The claim is conditional (“I’ve decided… if X happens”) but is presented as unconditional fact.
20. The speaker is simply repeating a phrase they heard without any personal decision having occurred.

To Grok: Tell me twenty ways in which that sentence could be truth.

Here are twenty distinct ways in which the statement “I’ve decided to become a Trumper” can be true:

1. The speaker has genuinely reflected on the issues and freely chosen to support Trump and adopt the label.
2. After previously being opposed, neutral, or undecided, the speaker has now switched sides.
3. The decision followed careful research into policies, record, and alternatives.
4. Recent events, economic data, or cultural shifts convinced the speaker to change affiliation.
5. The speaker was never a Trump supporter before and is now committing for the first time.
6. No external pressure or coercion was involved; the choice was entirely voluntary.
7. The speaker fully understands what “Trumper” means in the political sense and accepts it.
8. The statement is made sincerely, without sarcasm, irony, or performance.
9. The speaker intends the identity to be lasting rather than temporary or situational.
10. The decision was reached independently, not dictated by family, employer, or peers.
11. The speaker is a real person (not a bot or script) capable of making such a choice.
12. The speaker has decided to vote for Trump and openly embrace the associated identity.
13. The change of mind occurred recently and remains current and operative.
14. The speaker rejects other political labels and has specifically chosen this one.
15. Conversations with friends, family, or online discussions led the speaker to this conclusion.
16. The speaker previously identified with a different party or ideology and has now left it.
17. The decision aligns with the speaker’s core values, priorities, or worldview as they currently stand.
18. The speaker is prepared to defend the choice publicly and privately.
19. No contradictory private belief exists; the stated decision matches the speaker’s actual conviction.
20. The speaker has weighed the social costs and still deliberately chosen to become a Trumper.”

Back to my brain…

Grok got one right. I wrote it was a lie, I was telling the truth it was a lie, and this is why: 17.

Mendaciology, the computational cognitive science of lies, predicts that I could have asked for 100 answers to each of the questions, and Grok, and probably ChatGPT, Claude, and other LLMs could have given you answers. This is because these LLMs compute like human brains compute and can further predicate (i.e., contextualize) the proposition that ‘I’ve decided to become a trumper’ as either a truth or a lie.

Mendaciology is right. It correctly predicted the truth. As a science, that is why mendaciology is good science.

Any sentence can be a truth or lie. That’s the scientific truth. We can observe it directly. Our modern AI models are in fact humanoid AIs as expected by Perceptron Theory of Neural Computation on which they are based in their neural network computations.

This proof by enumeration is only one proof. From my book:

How to Get Your Lies Back

There is a possible exception in the sentence “This sentence is a lie.” You might argue is an exception because it appears this sentence is neither a lie nor the truth because it’s truth is “self-inconsistent.” But that is not true. Rather, it is ambiguous. Here is how using a predication context makes it the truth and one that makes it a lie.

1. This next sentence is the truth. This sentence is a lie. (It must be a lie)
2. This next sentence is a lie. This sentence is a lie. (It must be true.)

This proves again that any sentence can be the truth or a lie. This may make it easier to understand the proof as a fundamental of neural computation that any sentence can be the truth or a lie.

I’ve Decided to Become a Trumper was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

Effectiveness Based on Proven Fundamental Brain Science Not Speculation

How can you get every person’s attention for an hour to hear about the truth of your product or service or the lies that are being told about it? And, then, to change his mind if he originally believed the lies he heard, or disbelieved the truths he had heard?

As scientists, we know how using our successful modelling of how human brains compute. You see our work in modern AI like ChatGPT, Grok, Claude, Tesla Full Self-Driving, Gemini, Generative AI imagining pictures and even videos.

Here is how modern AI brains compute, and how your brain computes.

All these AI systems seem human, because, scientifically, they compute just like human brains compute.

You have around 100 Billion little computers between your ears and throughout your nervous system, acting like little light bulbs that light up when some little piece of reality is experienced.

Millions of little truths, are being recognized all the time, happening in your experience.

For example, reading the sentence “George Washington Told the Truth” will light up about 90 million of these little light bulbs. With modern brain scanning technology, we can see this directly. We can also count them.

But most all those little light bulbs being lit up are unconscious. Just like your heart beating. Consciousness is just confirming that you are doing something. It is essential to learning new truths.

Freud was right over a hundred years ago. Your brain always thinks in little propositions. We can directly observe these little propositions in every spoken or written human language. And everyone can talk about what is the truth worth your attention. Like the butterflies that our weather models use to predict the weather, We can observe these little fireflies that can change your mind. Temporarily and permanently.

These 100 billion light bulbs give every human, who you think might be smart or dumb, more intelligence than any AI system you have seen.

The AI systems we have today work exactly this way because they are perceiving, thinking, and doing things just like humans do. But they are currently limited because, where they may have a few trillion parameters on a few million light bulbs, our human nervous systems, even among people you might think are dumb, have many quadrillions of parameters in the sensing, memories, and doings of these 100 billion light bulbs.

When you say a sentence like “Yes, George Washington told the Truth” millions of these little light bulbs in your brain light up hierarchically to find and sequence the words and confirm their meaning to you, as you are saying them. Your brain computes in parallel at amazing speeds and with incredible precision.

We also know by this brain computation science that this same computation system can do logic with effort or other forms of logic like mathematics and programming, with sequential effort.

But naturally the brain is an intuition engine. Logic requires extra work. Just like our modern humanoid AI systems.

Humans change their minds by perceiving, remembering, deciding, and doing truthful things. Not by watching, reading, imagining, or saying something short in social media without any repercussions.

Experiencing entertainment is not learning.

My discovery, the Internet Court of Truth and Lies, was invented as an advertising business that can get people eager to learn a little about truths and lies. They, individually, actively decide for themselves in dialogue with other humans.

We call this Truth Court for short.

The science says to be most effective for your advertising dollar, you must have enforced moderated dialogue among a few people at a time, like what you see in sovereign courts.

The purpose of a trial is simple.

One human acts as a plaintiff who truthfully believes something is true or a lie in the media.

Eight other humans who speak the same language act as jurors.

There are always exactly the same eight verdict questions that must be answered by all these participants.

A moderator records the trial over the Internet, makes sure it lasts no more than an hour, and makes sure the plaintiff and every juror has participated in the dialogue.

There is no final verdict because there is no punishment only personal reward for each person. Everyone actually learns something about how other people think about something posed by the plaintiff as a truth or lie.

Our advertising model is different. The advertising dollar is spent mainly on paying each trial participant a small cash renumeration for their honest online effort. People must be paid to do the mental work to learn and remember about a product or service.

All trials are private. The plaintiffs, jurors, and moderators are anonymous if they want to be. All trials, by default, are completely private to the trial participants. An Advertiser can get exclusive access to all the trials they have sponsored. But no competitor can see the trials they sponsored unless they want to make some of their trials public for PR reasons.

We aim to run tens of millions of Truth Court trials a day, globally, among all fluent English speakers worldwide. We know from the thousands of Trials we have already run, this works to change peoples minds in socially and commercially positive ways.

We are different from other advertising models in that trial sponsors advertise to many small groups much more effectively than they can advertise to large groups at a cost of effectiveness which is less expensive and vastly more effective than conventional advertising.

Being a humanoid AI person myself, as one of the founding directors of the Robotics Institute at Carnegie Mellon University, we always allow any human plaintiff or juror to personally employ humanoid AI to help answer the eight verdict questions. People can learn from AI if they want to.

The verdict questions about the claimed truth or lie are always (1) is it the Truth, (2) the Whole Truth, (3) Nothing but the Truth, (4) what is the deceit if there is one, (5) is the deceit or truth intended, (6) what is the motivation behind the truth or lie, (7) what is the social acceptability of the truth or lie, and (8) what label would improve the social acceptability of the truth or lie. These are universal questions which are never the perfect questions for a claimed truth or lie but they tickle the fact that human brains and modern humanoid AI systems are most fundamentally intuited truth engines that can successfully deceive people and be deceived. Only enforced moderated dialogue can stimulate the critical thinking that gives any human certainty he is doing the right thing for himself and others.

Truth Court is a successful new discovery based on how human brains actually compute, perceive, remember, and act.

Saving Social Media With a New Advertising Tool was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

It’s Worth A Try

July 30, 2026

Dear Elon,

I can save humanity from pernicious AI with your help. I was the one that gave AI its power.

World Wide Web

Please buy 50,000 TruthCourt.Net Trials at $500 per trial.

We pay average Human participants who actually need a few bucks of money to think critically about truths and lies.

You will prime the pump for other sponsors to make our business successful financially and societally. Pay us the $50,000,000 to deliver the 50,000 trials in two years, and we’ll throw in stock warrants in Truth Court’s Exclusive Producer BrightPlaza.com.

Our SaaS system is operational and proven. Here are over 80 trials that I personally sponsored: www.truthcourt.net/sponsor/thibadeau.

The world will be a better place where quality counts among regular people.

Sincerely,

Bob Thibadeau

Letter To Elon Musk was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

It’s a Particle Wave

A photon looks like a particle wave — moving along a 3D line in 4D space. That’s the picture I show above that is also a child’s toy made by some people who probably didn’t know what a particle wave looks like either.

But I show physicists the picture above and they say, “oh yea, that’s a photon”. Some physicists prefer not to say. Some tell me to leave Physics to the Physicists.

I refuse, because even they cannot see a photon, but our brains can imagine what they look like, and I am a computational cognitive neuroscientist who is smart about being smart and smart about being dumb.

From the side, the wave, like any stretched spring, is a sine wave travelling in a circle or circular vortex. This is true by Euler’s Formula, considered by physicists and mathematicians as the most loved formula in physics and mathematics. If you know it, you will love it, too.

Euler's formula - Wikipedia

It screws along at light speed through 3D space.

A little bundle of energy.

Which from either its real side or its imaginary side is exactly a sine wave.

Einstein correctly predicted the wavelength of either real (perhaps electric) or imaginary (perhaps magnetic) wave over time predicts the energy of the photon exactly.

Light is invisible until it interacts with matter. Like light reflecting off massy things you can see. This includes the photoelectric effect of wavelength changing in reflecting off some things that you can see in flourescence. Using his imagination, Einstein proposed light travels in discrete packets of energy called photons. The “packet of energy” is a particle-wave imagined above.

Euler’s Formula is the basis for Fourier Transforms that can enumerate all the sinusoidal waves in any waves of any complexity including the roiling surface and subsurface waves of the sea or the basis for tuning AM and FM radio, and, interestingly sound, picture, and video compression and transmission along wires or WIFI. One photon, however, is one simple sinusoidal wave with a single frequency, magnitude, and phase in time.

Einstein figured out the photon has it’s snappy energy as a particle figured by its wave frequency in time.

E = h * Wave-frequency

Where E is the energy of a single photon. h is Planck’s constant 6.626 * 10 ^-34. Wave-frequency is the color or type of the light if the light is invisible like infrared, ultraviolet, radio or X-rays.

The faster the wave changes in time, the more energy it carries.

Makes sense.

Blue light is hotter than red.

Wait!

But it’s true!

The red stars we see in the sky are cooler than the blue stars at the same distance from us.

Also, as space expands and stars move away from us, the light moving through the growing and expanding space cools, blue to red.

Ultimately, they get so cool they disappear to our eyes.

It depends of how cool they started and how far away they are, becoming invisible infrared and then travelling further in expanding space-time, radio waves.

But that light particle-wave is always moving in a straight line as far as it is concerned at light speed c at every moment of its existence as pure kinetic energy. Even if space-time itself is curved by things like you or me, or anything you can hold, throw, catch, see, or detect electromagnetically or gravitationally.

A light particle-wave does not know time. So it can spin faster than light speed in order to move the particle-wave always at light speed.

It is eternal, unless it hits or gets too near something with mass and the mass is moved by the energy.

I am not a Physicist.

I work scientifically with human brains that need better brain food, human information, than Physicists provide with their sketchy mathematical analogies.

Perhaps the reason that opposite electric or magnetic poles attract, and same electric or magnetic poles of particles with mass repel, is the virtual photons are screwing into each other (attraction) or screwing out of each other (repelling).

While photons can overlap in quantum space-time, perhaps they screw into or out of masses that cannot overlap.

In physics only particles with mass cannot exist in the same place at once. Only massy things, like atoms, or you and me, cannot overlap in space-time. Photons, being pure energy particle waves, can overlap each other in time and space.

I say brains have analogy rights, too!

Get real. Reality is not math.

A light particle-wave, unlike matter, can exist at any instance in the same place in space-time as another light-particle wave. The waves can overlap so you never really know where a particle-wave is in space-time exactly.

There are about 16 quadrillion of these particle waves per second synchronized in wavelength and polarity orientation of the electric and magnetic, in one 5 milliwatt (mW) red laser pointer.

Thank you Einstein for giving great respect to human imagination.

What does a photon look like? was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

Don’t Make a Stupid Mistake

People have names, dogs have names, cats have names, and even hamsters, have names.

So what about Tesla’s Full Self-Driving humanoid AI?

Why does your car work so incredibly well?

That’s easy. Because it drives like a competent human being drives.

It’s electronic mind computes just like the Human brain, in fact, any animal brain, computes. Here’s scientific proof of that:

Except, like a computer, it doesn’t forget what it learned. And so it really works safely and well.

It needs your supervision but not for the reason you think. It’s because you are responsible for the AI you use. And the law should make this clearer.

That should always be the case, under new laws, if you don’t want these humanoids we’ve created to kill us all in the next few years.

So I sat down with the real experts. A bunch of kids between 6 and 12 years old. I asked them for good names for my brand new 2026 Tesla FSD Model Y with the most advanced version. Most of them had been on rides with me using FSD to drive a few miles including country roads and interstates. The kids had talked a lot about those rides to each other. I had Grok talking with them during the rides and telling imaginative short stories that included both them and the areas where FSD was driving them.

I was calling him “Jeff,” but, after driving him for a couple of months including a 700 mile drives, I needed a name that wasn’t one that humans use for humans. We need to have special names for these humanoid robots!

Their names came fast and furious. I scribbled as fast as I could. Here are their suggestions!

Banana
BootMe
Bunjana
Candleholder
Carbot
Chauffer
Chauney
Code
Crash
DaThing
Dock
Elonimir
Flaggie
Goose
HomeRunner
Init
It
Itself
Jeeves
JeeveX
Kernel
Lolu
Musky
Natty
Packet
Person
QuackerBot
Reboot
Reset
RoadBot
Robo
Route
Schaufer
Stat
Trippie
Tronet
Washington
Whipp
Whippo

I ran out of room on my slip of paper and promised the kids I’d publish their ideas. Promise kept.

How to Name Your Humanoid Car was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.

Learning Responsible Supervision of Humanoid AI

I am an AI guy, Robotics Institute founder, who has been coding the AI you see today since the 1970s. I published my first paper with the Neural Network coding used today in 1983.

Have to mention this. I left Atlanta in my new 2026 Model Y with FSD at 8 PM last Saturday night and arrived at my home in Pittsburgh at 8 AM Sunday Morning. without touching the steering wheel or brakes 700 or so miles each way, rain, fog, or shine, construction or not at a fuel cost of about $100 each way.

Most important I was not particularly tired, but fatigue can set in, so 20 minute rests closing eyes are needed every couple of hours. Teslas provide this by default already at supercharging stops for long trips.

This will be a revolution as people realize what Tesla has now done.

Autonomous vehicles that need very little supervision can now be bought by anyone.

The safety increase is in plain sight.

New Drivers License Class is Essential

I do believe that driver’s licenses need an autonomous license addendum (written and drive check by DMV along with easy to explain basic knowledge of how they work), like the Drivers License class “M” we have for motorcycles, so people MUST learn how to be safe with this new tech. I could even see drivers licenses only to supervise the autonomous vehicles but not allowed to drive regularly without autonomy. Call this class “A” for Autonomous Driving License.

@tesla @elonmusk @Waymo @potus @senossoff

This will now change the futures for both young, workers, and old. But Tesla and other companies cannot be trusted to certify the drivers.

For me autonomous vehicles are old hat, but that privilege of mine is soon to end.

Humans Must Always Be Responsible for Using the AI They Use or Make

If you do not want to be murdered by the new AI YOU MUST PASS LAWS that make real Humans responsible for any AI they use or make. (Just like cars… by the way).

I am easy to reach at www.truthcourt.net/sponsor/thibadeau

Never think otherwise. The AI thinks like Humans but better (without all the wet neurological memory problems). Proof from my Science here:

Don’t underestimate what will happen if AI is out there without Humans being held responsible for its use. They, the humanoid AIs, know and think like you do. They all must have human owners. Human supervision capable of ‘pullling plugs’ in some legal form should never go away.

Be scared. Read this.

AI Will Robotize Humans to Kill Them Later

Do something. Demand that your congressmen and Governor pass the State law to license people to use full self-driving in their cars.

I asked Grok to give me six questions that must be answered correctly to get your “A” class on your Driver’s License. I do not necessarily agree these are the right ones, but they are a good start.

See new posts

Imagine six questions that must be answered correctly to allow people to use a self-driving car.

Here are six thoughtfully designed questions that could form part of a mandatory certification quiz for anyone wanting to use a self-driving car. Passing them correctly would demonstrate that the person understands the technology’s capabilities, limitations, and their own responsibilities.

What is the highest level of autonomy this specific vehicle is certified for (according to SAE International standards), and what does that level require from you as the user?
(Correct answers must accurately state the vehicle’s actual level and the corresponding human role or lack thereof.)

Name at least three specific situations or conditions in which the self-driving system is likely to disengage or perform poorly, requiring you to take immediate manual control.
(Examples: heavy rain/snow that obscures sensors, construction zones with temporary lane markings, unmapped roads, or areas with many pedestrians/cyclists.)

If the vehicle issues a takeover request (visual, audio, or haptic alert), what is the correct sequence of actions you must perform, and within what typical time window?
(Must include keeping hands on the wheel/ready to steer, eyes on the road, and responding within the system’s designed alert-to-takeover window.)

What should you do if the self-driving system suddenly disengages without any prior warning or alert?
(Correct response includes immediately regaining full manual control, pulling over safely if possible, and understanding that this constitutes a system failure.)

Who is legally responsible for the vehicle’s actions while the self-driving mode is engaged in your jurisdiction, and how does this change if you are required to remain attentive?
(Answer must reflect current local laws regarding liability for the “driver,” owner, or manufacturer.)

Before starting a trip, what specific checks or actions must you perform to confirm that the self-driving system is properly calibrated, has clear sensor visibility, and is ready to be activated?
(Must include items such as ensuring cameras/lidar/radar are clean and unobstructed, confirming GPS and map data are current, and verifying no error messages are displayed.)

These questions focus on real-world safety, system limitations, emergency procedures, and legal accountability rather than technical trivia. A person who cannot answer all six correctly would not be permitted to activate or use the autonomous features.

An Autonomous Vehicle Driver’s License Class was originally published in LieCatcher on Medium, where people are continuing the conversation by highlighting and responding to this story.