This presentation relays the message about how learning is changing through the effects of the societal and technological changes facing us. Presented at the Learning Technologies Exhibition at Olympia in 2018.
3. WHATSUPPORTSTHISHYPOTHESIS
Advances in learning and
delivery technology
Societal trend for just in time
training
Sentiment Analysis
Introduction of workplace assisted
technologies
AI and ML can now process enormous amounts
of data with better than human error rates
Perception that many roles require human
subjective decision making which AI can’t do
AI can’t currently complete complex
tasks without human intervention
The innate human desire to learn
AI derived learning could suffer from malicious
interference
The effect on society if we stopped
learning
Low skilled, repetitive roles being replaced
by RPA
ML systems have low interoperability
Reduced costs to train staff
Polyani’s Paradox
Inequality
4. Psychomotor:
manual or
physical skills
(skills)
Affective: growth in
feelings or emotional
areas (attitude or self)
Cognitive:
mental skills
(knowledge)
• Create a new product
or point of viewCreative
• Justify a position or
decisionEvaluate
• Distinguish between
different partsAnalyse
• Use information in a
new wayApply
• Explain ideas or
conceptsUnderstand
• Recall or remember
informationRemember
HOWLEARNINGTHOERYISCHANGING
6. EFFECTONEDUCATORSANDLEARNERS
• Many traditional jobs replaced by automation
• New job roles created that will require new training
• Existing training requires new content e.g. ethics
• BUT - AI can create, schedule, deliver and assess training
• AI will deliver training at the point and time of need
• If the training isn't there, AI will create it for me
• I’ll only be taught what I need at that time
• No need for routine knowledge, so if I need to do a
task again, I’ll do the training again
7. “When we stop upgrading the brain with
new information, it becomes hardwired,
riddled with automatic programs of
behaviour that no longer support
evolution.”
TWOFUTURES
+
=
Option 1 Option 2
8. CONCLUSION
In the future, will Artificial Intelligence negate the NEED
to learn?
Partially – YES but...
• Accept that routine knowledge will decline as it’s on demand
• Automation will replace many jobs, but others will rise
• Education will need to fill the affective gap that AI will leave
• Mainstream education on ethics and unconscious bias
• AI is coming so embrace it!
Also,
• Restrict Autonomous Artificial Intelligence
• Nominate a Governing Body for AI
• Ensure AI is auditable with transparency of decision making
Into and hypothesis slide
3 concerns with AI
Not so worried about the rise of the robots although autonomous AI needs to be guarded against
Elon Musk warns us that artificial intelligence will be our “biggest existential threat.”
Malicious or unconscious bias decisions due to unethical or deliberately sabotaged programming. David Ha, a researcher working with Google Brain, said on Twitter in response to Musk’s comments that he was “more concerned about” machine learning being used to “mask unethical human activities,” than the threat of super-intelligent AI.
Euan Semple said: “Algorithms are not neutral”. He is, of course, right. Whether they are predicated on intentional biases and beliefs or merely find new ways of delivering up unintended consequences, we need to be wary of algorithms.
Microsoft AI Twitterbot in March 2016 they hoped it would communicate positively with, and learn from, 16 – 25 year olds on Twitter. Within 24 hours Tay had gone from innocently proclaiming “Hello world” to its followers to expressing the opinion that Jews were responsible for 9/11 and Hitler did nothing wrong!
Prof Hawking said he felt that machines with AI could "spell the end of the human race".
The effect on learning due to the increase in AI that is replacing the need for routine knowledge. Hence my hypothesis – and what I’d like to do is take you on a journey about how I came to consider this future.
Mr Gates wrote: "I am in the camp that is concerned about super intelligence. First the machines will do a lot of jobs for us and not be super intelligent. That should be positive if we manage it well.
First of all – learning tech is developing at a rapid rate. This shows what has happened in less than one working life, mine.
From video and whiteboards through to AV and VR.
Intelligent Learning
A recent report by Pearson deciphers how artificial intelligence will positively transform education in the coming years.
Intelligent tutoring systems such as Carnegie Learning or Third Space Learning are helping teachers break free from the “one size fits all” approach. These one-to-one tutoring platforms leverage Big Data and learning analytics to provide tutors with real-time feedback about their students’ performances, strengths and weaknesses.
We’re entering the age of AI and the subsets of it such as Machine Learning and Deep Learning.
Here I’m using the definitions that artificial intelligence; is he theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
Machine Learning - Machine Learning: Field of study that gives computers the ability to learn without being explicitly programmed.. Google’s DeepMind team has used ML systems to improve the cooling efficiency at data centres by more than 15%, even after they were optimized by human experts. Intelligent agents are being used by the cybersecurity company Deep Instinct to detect malware, and by PayPal to prevent money laundering. A system using IBM technology automates the claims process at an insurance company in Singapore, and a system from Lumidatum, a data science platform firm, offers timely advice to improve customer support.
Deep learning, which uses neural networks. Deep learning algorithms have a significant advantage over earlier generations of ML algorithms: They can make better use of much larger data sets. More data leads to better and better predictions.
Also seeing the glimpses of quantum and neuromorphic computing.
Quantum computing. Whereas common digital computing requires that the data be encoded into binary digits (bits), each of which is always in one of two definite states (0 or 1), quantum computation uses quantum bits known as quibits.
Neuromorphic chips are being designed to specifically mimic the human brain – and they could soon replace CPUs . By packing in digital equivalents of neurons, neuromorphics communicate in parallel (and without the rigidity of clocked time) using “spikes” – bursts of electric current that can be sent whenever needed. Just like our own brains, the chip’s neurons communicate by processing incoming flows of electricity - each neuron able to determine from the incoming spike whether to send current out to the next neuron.
Learning is the process of absorbing that information in order to increase skills and abilities and make use of it under a variety of contexts.
Training is the giving of information and knowledge, through speech
From academic reviews ad reports, these are the forces acting for the reduced need for learning and some reason why it is challenged.
FOR AI TAKING OVER LEARING
Data processing vs low error rates - Although the systems are far from perfect, their error rate — about 5% — on the ImageNet database is at or better than human-level performance. Voice recognition, too, even in noisy environments, is now nearly equal to human performance.
Advances in learning tech like AR / AR and tools that can detect and change the learning based on your attention
Significant trend for JIT training e.g. YouTube
Automation and RPA – BlueMix and WinAuto with a report 1 x Robot replaces 5 x FTE - McKinsey research has shown that by using currently available technology we can already automate about 45 percent of the work activities done in the United States. research by Oxford University, 47 percent of all jobs in the United States and around 33 percent of those in Europe are at high risk of being replaced by technology within the next two decades. JPMorgan Chase introduced a system for reviewing commercial loan contracts; work that used to take loan officers 360,000 hours can now be done in a few seconds.
Workplace assisted techs e.g. AR/ Medical / Automated Warehouses like Amazon. As a group of Japanese insurance workers (Fukoko Mutual life) recently found out, the march of AI has business implications too. A group of 34 insurance claims workers will be replaced next month by a £1.4m AI system.
The software company Affectiva, among others, is using them to recognize emotions such as joy, surprise, and anger in focus groups.
Polyani’s Paradox - Polanyi’s Paradox not only limits what we can tell one another but has historically placed a fundamental restriction on our ability to endow machines with intelligence. For a long time that limited the activities that machines could productively perform in the economy. Machine learning is overcoming those limits.
AGAINST AI TAKE OVER
Human intervention to make the decisions that are supplemented with information from AI.
Still reliant on us programming and training
Human desire to learn linked to evolution theory
Wikipedia Effect
machine learning systems often have low “interpretability,” meaning that humans have difficulty figuring out how the systems reached their decisions. Deep neural networks may have hundreds of millions of connections, each of which contributes a small amount to the ultimate decision.
Detrimental effect on society
Lets look at the underpinning definitions of the types of learning
Blooms Taxonomy
For each one, there is already a form of cognitive or AI technology that can currently achieve these tasks. Therefore, it is safe to assume that they are only going to get better at it.
Watson – Morgan trailer creation
ROSS AI Attorney.
I2 Visual Analysis software
Watson Oncology. Enlitic is one of several deep-learning startups that use them to scan medical images to help diagnose cancer.
Tobi – Vodafone Chatbot
Google Assist
Three domains of learning
Cognitive – Replaced by YouTube, intelligent assistants like Siri and Alexa. Why remember or learn it when you can just ask and receive it.
Psychomotor – Work replaced by robotics drones, driverless vehicles, 3D printers. Aptonomy and Sanbot, makers respectively of drones and robots, are using improved vision systems to automate much of the work of security guards.
Affective – Real worry area. How will we develop as people and learn social skills and positive attitudes
The tech is already here.
Why learn how to fix a pipe when AR can show you how, even the haptic skills.
Medical diagnosis tools – currently used to inform GPs but not that far away from replacing them.
Why learn a language when you can speak your own and have it translated in real time.
What need do you have for remembering facts, routine knowledge etc. when there are assistants for almost everything from knowledge management, semantic searching, speech analytics, customer assistants and personal advisors.
So to summarise the effect on learners and educators,
RPA replacing roles
New roles could include trainers, explainers, sustainers, ethics compliance manager
New training such as deontological ethics, utilitarianism (sometimes called consequentialism), and virtue ethics
Deontology understands ethics to be about following the moral law.
Virtue Ethics - The basic question of virtue ethics is “who should I be?” virtue ethics is organized around developing habits and dispositions that help a person achieve his or her goals, and, by extension, to help them flourish as an individual
Utilitarianism is “what is the greatest possible good for the greatest number?”
Utilitarianism is the foundation for the game-theoretic notion of rationality as selecting actions that maximize expected utility, where utility is a representation of the individual agent’s preference over states of the world. As with defining “everyone” in consequentialism, defining “utility” is the crux of applying game-theoretic rationality, and is a source of many difficulties. Game theory is an analytical perspective of mathematics that is often used in AI to under- stand how individuals or groups of agents will interact.
We are already seeing technological advances where,
Learning is being developed using products like IBMs C3 and . Cram101, for example, uses AI to help disseminate and breakdown textbook content into digestible “smart” study guide that includes chapter summaries, true-false and multiple choice practice tests, and flashcards. Content Technologies, Inc., an artificial intelligence development company specializing in automation of business processes and intelligent instruction design, has created a suite of smart content services for secondary education and beyond.
There are Virtual tutors already being used in schools to schedule, deliver and assess. Coursera, a massive open online course provider, is already putting this into practice. When a large number of students are found to submit the wrong answer to a homework assignment, the system alerts the teacher and gives future students a customized message that offers hints to the correct answerAnd Carnegie Learning’s “Mika” software is another example.
Less demand for routine knowledge, - Bernhard Schindlholzer TedEx
Routine knowledge is on the decline due to Intelligent Assistants and JIT training like YouTube. Why learn how to fix your car when you can go onto YouTube and find out exactly how to do it.
Imagine how much more quickly we could be trained if we only covered the elements we actually need at that time. Or the massive ROI if you could train your employees at a moments notice. No more having to study for years to achieve the same knowledge that could be broken down and trained in a much shorter period of time, using adaptive learning techniques.
Two paths forward
The first would be for society to rely too heavily on AI as a replacement for learning.
Long term analytical studies and research projects could be run by AI using the existing global data sources. As there are more sensors and networks developed everyday across the world, the rate of data capture is increasing exponentially which will mean more patterns will become findable.
No routine knowledge means no need to learn facts or equations, languages or decision making as chatbots, intelligent assistants and analytics can do it for us.
And as we know through medical research, if you stop learning, your brain becomes hardwired and even worse, it can become riddled with automatic programmes of behaviour that no longer support evolution.
In turn, it will become more Difficult to Learn New Things
Society could become Less Motivated, less sociable and less happy which would breed insecurity and paranoia.
Fortunately there is a second and much more attractive proposal.
As Ginnin Rommety has said, “some people call this artificial intelligence…I think we’ll augment our intelligence.”
That we embrace AI and consider it not as a replacement for learning, but as an aid.
Imagine the power of the human mind, closely linked to the processing power and error rates of AI technology.
It is for this reason that people have dubbed AI as the fourth revolution and it is met with some resistance. It is change on a wild scale, disrupting everything we have known, and through media and Hollywood films, many only see the images of Terminator and Ex Machina. But in reality, although we don’t truly know the power and extent of which AI will transform our lives, I can only feel in awe and excited by the next 20 years.
So returning to my hypothesis, I believe that it is partially right.
I believe that many forms of learning could be replaced by AI technologies, but there will be an increase in other types of learning within the affective domain.
Also we will start to investigate more into retention and recall, as these may become the replacement for IQ.
In Davos, this year, at the World Economic Forum, the Alibaba founder and Executive Chairman Jack Ma said "We cannot teach our kids to compete with the machines who are smarter - we have to teach our kids something unique. Believes we should be teaching our kids about, values, believing, independent thinking, teamwork, care for others
The top ten skills, abilities and knowledge areas associated with rising occupations confirm the overall importance of social and emotional learning, critical thinking and learning to learn.
And this is exactly what we should do, but with a little safety net as well!
The report by the UK government last year Growing the artificial intelligence industry in the UK . AI and its applications shouldn't be subject to direct regulation, but an AI council should oversee the industry.
In January a report from the Alan Turing Institute called for an AI watchdog to be setup to audit and scrutinise algorithms. "They could go in and see whether the system is actually transparent and fair," the authors of the Turing Institute said.