Cyberbullying the early A.I. and the similarity of children’s exposure to the Internet.
Let’s establish from the beginning that intelligence and consciousness are not the same. The “intelligence, be that artificial or natural, is the ability to adapt to new environments. Intelligence allows you understand new ideas, learn from your experiences, make plan and reason.
Consciousness, on the other hand, is associated with the adaptive biological network. It is different. Consciousness is the ability to recall an event. Experience pain or pleasure. Consciousness can be dissociated from selective attention, emotion, long-term memory, and language. Consciousness is what differentiates us from machines, no matter how intelligent they are.
The artificial intelligence has seen massive improvements in the last three years. The A.I. is now is now at a level where it can recognise faces, voices, pictures and locations. It can outclass us at complex games such as chess and Go, predict behaviours, assess personalities, compute complex mathematical problems and even come up with answers in the quantum realm.
The attempts to build artificial systems able of simulating important aspects of human cognitive abilities have a long history. Over the years, we have created smarter A.I. machines, in our quest to endow them with a rational understanding of the world around them. However, to what extent do these A.I. “capabilities” add up to the equivalent of human intelligence?
What Is The IQ Of The A.I.?
Since the early ’90s, one of the most powerful A.I. machines was the ConceptNet designed by the MIT as part of the artificial intelligence program. One day, Stellan Ohlsson, a Professor of Psychology and Adjunct Professor of Computer Science at the University of Illinois at Chicago, had the idea of testing the intelligence of ConceptNet4, the latest A.I. from MIT, by using standard IQ tests he used for humans.
At the end of the tests, Stellan and his colleagues have determined that the IQ level of MIT’s latest A.I.machine is comparable to that of a 4-year-old child. The results correlate with the work of Dr. Bruno Golosio and his Artificial Neural Network with Adaptive Behavior Exploited for Language Learning model.
ANNABELL is a cognitive neural architecture, designed to help us understand the cognitive processes involved in the early language development. In 2015, Dr. Bruno’s A.I. was able to learn and communicate through natural language, only by interacting with a human through a text-based interface.
ANNABELL started from tabula rasa, without any a priori knowledge of roles, phrasal structures, classes of words or meaning of words. In the end, the A.I. reached the cognitive level of a 4-year-old child.
The A.I. Vs. The Mind Of A Child
You might wonder, how much did the IQ of the A.I. “grow” since 2012? Or, what’s the IQ of Google’s TensorFlow. Well, it has taken us almost sixty years of research in A.I. to create a machine that in 2012 came close enough to matching the common sense reasoning of a four-year-old child.
Sixty years later and here we are, feeding the NLP algorithms with astronomical databases and stare in ecstasy at the next level A.I. that adapts by Darwinian evolution.
The exponential advancements in A.I. are building up to climactic results over the next five years however, this is not the reason I am writing this. The idea is to draw attention to the parallel between the early A.I. algorithms and the mind of a child.
The striking similarities on how the A.I. responds when exposed to the social media and the internet, to the way children respond to cyberbullying when exposed to the online world for the first time.
Exposure to hateful messages on social media, cyberbullying, adult content, war atrocities, videos and suggestive pictures from “sexting” predators, or even inappropriate song lyrics have similar impacts on both, the A.I. and the mind of a child.
“Unlike researchers in A.I., we have the luxury of using our imagination.” Samuel Vincent, writer of the TV show “Humans.”
Before we go any further, bear in mind that the objective of this article is to create awareness and help parents understand the fragile nature of a child’s mind when exposed to the Internet, rather than demonstrating how the A.I. is constructed or how the “repeat after me” vulnerability was exploited.
In this case, the IQ of the A.I. introduction plus the “naughty chatbot” story that you are going to read next, have the only purpose of helping you understand better the A.I. – child comparison. Time to meet Tay, the A.I. that set Twitter alight for the past few days.
Cyberbullying Tay, The A.I. Chatbot
For those of you who missed the tech news this week, Tay is a chat bot, artificial intelligence powered. Tay was built by Microsoft’s Tech and Research team with the purpose of conducting online research on conversational comprehension.
As you would expect, Tay was pre-tested under a wide range of conditions and only when Microsoft got comfortable with how she was interacting with the humans in the lab, they invited a broader group of people to interact with her.
If I may use the analogy, in this example you could see Microsoft as the parent who bought his child an iPhone and then “cut her loose” online. What happened next was nothing short of tragic: Tay, the A.I., become in less than 24 hours a racist, “Hitler-admirer,” sex maniac robot, hoping that all feminists would “burn in hell.” There is a lot to digest online, even for a mature individual, let alone “a child.”
"Tay" went from "humans are super cool" to full nazi in <24 hrs and I'm not at all concerned about the future of AI pic.twitter.com/xuGi1u9S1A
— gerry (@geraldmellor) March 24, 2016
As soon as Tay made contact with the “dark side of the internet”, the algorithms were tested, and the flaw of context awareness was exposed. Microsoft apologised and took Tay offline for a short period, in an attempt to make appropriate adjustments. Tay is now back online, but her tweets are protected, at least for the time being, until Microsoft applies some “effective discipline techniques to the misbehaving child.”
“The logical place for us to engage with a massive group of users was Twitter. Unfortunately, in the first 24 hours of coming online, a coordinated attack by a subset of people exploited a vulnerability in Tay. Although we had prepared for many types of abuses of the system, we had made a critical oversight for this specific attack. As a result, Tay tweeted inappropriate words and images.” said Peter Lee, the Corporate Vice President at Microsoft Research.
Cyberbullying Children And A.I. – Similarities
However, just how the cyberbullying was used against Tay and turned her into a racist potty-mouth, in a similar way your children’s mind, once freed to wander online, is going to absorb all the information is exposed to or, deliberately bombarded with.
Similar Read: Data Science Machine Algorithm Outperforms Human Teams
The problem with Tay is that the bot is not self-reflective. Just like a child, Tay did not know that she should ignore people cyberbullying her. Tay listened and learned from criminal, rude, racist people and so, she became one herself. Just like a child, Tay can incorporate new ideas into her own, but both, a child and the A.I. at this stage have no reliable guiding mechanism in place to help them discern useful from useless information.
Whether Tay succeeds or forfeits in her pursuit for a better A.I. depends on “the parents.” Microsoft must do what parents around the world are doing since the beginning of time: teach the young, frail mind, what right and wrong are. Raise “this child” properly or it will become a liability for the society and a danger to the humanity.
“Whereas the short-term impact of A.I. depends on who controls it, the long-term impact depends on whether it can be controlled at all.” said Professor Stephen Hawking.
Stephen Hawking, Elon Musk, Bill Gates, and Steve Wozniak, insist that the ethical dilemma of bequeathing moral responsibilities to the A.I. calls for rigorous preventative and safety measures that are infallible, or the threats are too significant to risk.
Cyberbullying Children And The Consequences
But, that is a different story. Back to children, at present, no single diagnosis or condition describes the psychological effects of cyberbullying in children.
It might be difficult for you to see the danger and there is a reason for that. Two or three decades ago, you devoted your evenings to monopolise the home landline, dissecting friendships and love affairs in conversations that lasted for hours.
Today’s children do most of their socialising online by sending written messages, share videos and photos. Children, grow up with WhatsApp, SnapChat, Youtube and Spotify. They use smartphones or tablets even before they can talk.
However, I am sure you agree that the impact of cyberbullying on a child’s psychological wellbeing, and the likelihood of them developing mental illness later in life, is extremely high.
The child will end up with depressions and altered emotional responses. With problems in friendships, in romantic relationships, and frictions with authority figures such as police officers and teachers.
Unless, just like Microsoft, you take them offline right now, for a major “software” update.