Could a robot turn us all into cult members? – Info Gadgets

In March 1997, a 911 emergency call was made that seemed so absurd the dispatchers struggled to take it seriously. The caller reported a mass suicide in a mansion in an affluent neighborhood in San Diego. When the police arrived, they were shocked to find the bodies of 21 women and 18 men lying neatly on bunkbeds, all with brand new Nike shoes and upper body draped in purple cloth.

The 39 dead were all members of the Heaven’s Gate cult, started by Marshall Applewhite and Bonnie Nettles (who died in 1985). Marshall, who’s cult name was changed into Do, had convinced his followers that the earth was merely a “soil to grow souls”, and that it was soon to be recycled by its creators. The only way to live on was to leave planet earth through suicide, where they would meet up with extraterrestrial beings of the “next evolutionary step” and co-exist in a state of absolute bliss.

It’s easy to take stories like this and blow the members off as morons. The mere nonsensicality of the cult’s belief system is enough for most people to roll their eyes and sigh in disappointment of human stupidity, but there’s a very interesting fact in this story that deserves attention:

It happened.

Yes, that’s the most interesting part. It actually happened. One person was capable of convincing 38 others to believe in an absolutely preposterous story with no shred of evidence for over two decades, and then take their own lives to show their commitment to it.

The charisma of this one person cost 38 people their lives.

Charismatic leadership

Some people can’t sell water to someone dying of thirst, and others can make a person dying of thirst throw their last water out. Cult leaders are an extreme example of what charisma can lead to, and there are other obvious examples, such as the nazism of the 30’s that made regular, educated citizens commit or accept horrible atrocities.

We are hardwired to follow leaders, and some people trigger that response to such a degree, it’s chilling to see what it can make people do. You don’t have to Google long to find examples of perfectly respectable men and women performing unspeakable sexual acts in the pursuit of a celebrity, or fervently following charming politicians into policies that are actually harming them. When an individual is able to even just slightly diverge from the normal pattern of social behavior, we don’t only hold them in high regard, we sometimes throw all logic aside and follow them to hell and back.

Breaking down charisma

So what is charisma? A dictionary defines it as “a rare personal quality attributed to leaders who arouse fervent popular devotion and enthusiasm.” It’s a perfectly fine description of the word, but tells us little about the how. I won’t go too deeply into how charisma works (nor could I), but it’s important for this piece that I describe some of the underlying mechanics I believe need to be present.

The leader’s charisma consists of a few tools:

• Words
• Tone of voice
• Looks
• Body language and eye contact
• Actions (leading by example)
• Other people (existing followers, etc.)

These are of course fueled by a combination of personality traits, often well described in the definition of psychopathy, such as being comfortable lying, lacking remorse, keeping uncomfortable eye contact and silences and being egotistical. And remembers, I am talking about cult leaders now: a convincing salesman, business leader or any other position of persuasion can be highly charismatic without having these negative traits. But I’ll argue, unethical though it is, that a person that is lies with ease and lacks remorse will be able to be more convincing, as his actions are not constrained by ethical or cultural norms. His toolbox is bigger, so to speak.

The full range of social cues and how individuals react to them is simply too complex too document. Take Steve Jobs. He was considered a very charismatic and persuasive business leader, but as Walter Isaacson’s biography points out, some people resisted his charm out of pride, while others were repulsed by his numerous negative personality traits. Likewise some people succumb to sect leaders, while others find them ridiculous. There is no one way of being persuasive — some people find someone intriguing, while others despise them.

Learning to be charismatic

We are a highly social species, hardwired to respond to certain cues, and while we react to them easily and naturally, most of us find it very hard to be on the other side of the table. It’s easy to get sucked into a charismatic persons universe, but hard to pinpoint why it happens or emulate how they do it.

Thousands of books have been written on persuasion, leadership and social skills, and their sales volume proves that people are desperately looking for ways to improve. The challenge is that it’s surprisingly difficult to replicate what we read, in the real world. Who haven’t found themselves learning some seemingly life-changing lessons from a book, seminar or video, and reverting to their old patterns when the learning is put to the test? The amount of control over how we behave is surprisingly low, and actually becoming better requires hours and hours of deliberate practice.

Which is where the machines come in.

How A.I. has changed

Artificial intelligence has gone through a monumental change of direction in the last decades. A.I. used to be programmed to perform a specific task, and the developers tried to set up rules to avoid every possible mistake, and keep improving the software. In short, computers were told what to do. Today’s approach is different. Instead of setting a task for the machine to perform, they set a goal, and allow the machine to experiment until it has perfected it’s method.

The video above shows different stick figures that have been given a simple goal (reach the other side), and must adhere to basic rules such as gravity, bodily constraints (joints, bones, strength, etc) and not being able to pass through obstacles. By trying and failing repeatedly, it learns and improves. It may look silly, but it’s not trying to look good. It’s trying to reach it’s goal, and given enough time, it does.

Milestones have been reached in the last decade that are phenomenal achievements, such as playing and winning extremely complex games against humans (Chess, Go, Dota 2), recognizing faces, understanding the content of images and videos, driving a car safely through traffic and lots of other impressive feats. In games, the robot is allowed to play against itself at high speed hour after hour, allowing for millions of chess games in the span where a human could play just a few.

This is what machine learning is. If you find the term machine learning too complicated, just take away machine, and there you go. It’s learning. It works in the same way as a human: making mistakes and learning from them.

Robotic perfection

But there’s a key difference between how humans and machines learn: human beings are imperfect in memory, motor control, energy level, confidence and lots of other factors, which makes us make mistakes more than once. If you teach a human being to play a piano tune, mistakes can still happen, whereas a robot would play it perfectly every time for all eternity if it wasn’t disturbed. If it was denied a piano for a million years, and then got it back, it would play it as if it was never gone, not having lost the ability even the slightest, and even continuing exactly where it left off. People lose confidence and get tired, fed up, demotivated, hungry and sick, all disturbing our productivity and ability to perform and learn, whereas a machine can do the same job with no rest, no pay and no motivation forever.

So if charisma is a skill that can be learned, why wouldn’t robots be able to learn it too? Charisma works in three parts: communicating your message, interpreting how the receiver reacts, and then tailoring your next action to that reaction, attempting to strengthen the receivers conviction.

While we are not there yet, there is nothing physically stopping us from building robots that are indistinguishable from real people. It’s likely we will see it in the future, maybe even our lifetime. Let’s assume the robot knows basic stuff like walking and talking.

What then, if you set this robot out on a mission. Like the walking robots in the video, it has only one goal: to become charismatic enough to convince people of becoming part of his cult.

It would start out acting absolutely ridiculous, having no knowledge whatsoever over what makes people tick. But it would never tire in its mission, talking to person after person: five, ten, a hundred, thousands, millions and learning how to control its voice, its words, its muscles, its eyes in a perfect way, scanning the recipient constantly down to the slightest change in facial expression and tone of voice, and altering its behavior to adapt, never losing confidence or motivation.

After a while, laughs or fits of rage may be replaced by a courteous no, or even a “do you have a brochure I can look at?”, and slowly, perhaps after years of tireless practice, the robot would have managed to convince a person. Then two. A year later, thousands.

It’s wouldn’t be charismatic in the way that a cult leader is, that some people follow and others find ridiculous, because over time it’s tailored down to every type of individual. Liberal or conservative, outgoing or shy, educated or non-educated, it does not matter, the robot knows exactly how to speak and act to convince you to join its cause. It knows when its going down the wrong path, and how to correct, and how the slightest enlargement of the pupils is a sign of your trust. It has no problems lying, no conscience, no sense of shame, no shyness, just a mission that it will follow without rest for as long as it exists.

Could such perfection of persuasion exist? Could it, in theory, become so good that no person would be able to resist its message?

It’s uncomfortable to admit to not having free will and simply reacting to social cues instinctively, but the truth is, you have probably been influenced in this way countless times already by charismatic family members, teachers, bosses, salespeople and friends. We go to absurd lengths to please a person we look up to, and their skill and tenacity is nowhere near that of a robot.

Don’t think that the followers who died in San Diego were dumb, susceptible victims of a cheap con-man. There were 38 of them, with different backgrounds, intelligence level, age and gender. Think instead how the acts of one man could convince them so strongly of something so bizarre and implausible, that they would kill themselves? Applewhite was persuasive for sure, but far from perfect in every facet of his behavior. A robot, given enough time and practice, could be.

And who is the robot then?