L&D 2020: Shaping change in learning

Technology

Artificial intelligence

Artificial intelligence (AI) has become integrated with e-learning¹ to produce ‘intelligent e-learning’ (IEL) systems. IEL has now progressed significantly beyond early AI systems, to teach a language for example². Now combined with AI, e-learning is no longer simply a passive exchange of learning between user and computer. Instead, the individual’s learning is actively mediated by an avatar – the AI part of the system. AIs have infinite patience to help learners and provide detailed feedback³ as learning progresses.

Current programmes provide high quality virtual reality that simulates real-life situations – for train drivers, construction workers, riggers and production line employees, for example. This type of learning has been particularly beneficial for hard-to-reach groups where attendance at traditional face-to-face events has been expensive and difficult to arrange.

AIs assess a learner’s strengths and weaknesses as they progress through a programme. As a result, the AI adapts the programme through an understanding of the learner’s capabilities. The AI, represented by an avatar, recognises and adapts to the learner’s difficulties and emotional distress. The avatar knows when the learner is struggling and provides further learning programmes and encouragement. As a result, learning retention is now more than 40 percent above many traditional human-based programmes.

Companies are also using AI programmes for corporate communication – to communicate changes in strategy, for example. In the case of global corporations, the AI instantly translates any text and even ‘learns’ dialects to make the delivery more locally relevant. These systems also provide a real-time feedback channel.

Most AI systems are based on earlier developments at the MIT Media Lab⁴ that gave computers ‘common sense’ – the ability to understand and reason about the world as humans do. In this way, devices can understand people’s goals and typical problems and computers now regularly assist executives in complex problem-solving. These intuitive computerised devices have become simply more ‘human’.

Computer intelligence is so superior that, to make the most of the human/computer interface, individuals are starting to enhance their brain functions with neural implants. Without these implants learning is just too slow. This enhancement had already been predicted by Kurzweil⁴ who said: ‘Once machines become as intelligent as people, they will necessarily soar past us.’ He added that computers evolve faster, remember more accurately and share information more efficiently than non-augmented humans can.

The amount of available knowledge is increasing exponentially, but computer knowledge has the key advantage of being directly downloadable from one computer to another. Although direct person-to-person transfer is now possible, using AIs to compute the information produces a far better synthesis than would be capable by human minds alone – even with direct linking.

Work is also progressing on transferring human minds to computers, by exact copying. The resulting AI, sometimes called an ‘infomorph’⁶ generates responses indistinguishable from the original organic brain. Experiments are underway to place the downloaded AI into a confining artificial body – an intelligent robot. The Tyrell Corporation is planning to use such robots as trainers in hostile situations such as battle zones, mining, space and deep sea environments.

AIs are regularly combined with human intelligence to produce a higher intelligence than either one could achieve alone⁷.

1. http://www.canada.com/montrealgazette/story.html?id=263e8270-7170-4588-a5ac-3a731d0c99e6 04/07

2. http://www.usc.edu/uscnews/stories/10321.html 06/04

3. http://www.aaai.org/AITopics/html/tutor.html

4. http://csc.media.mit.edu/

5. http://www.eetimes.com/special/special_issues/1998/timespeople98/kurzweil.html

6. http://www.gurpswiki.net/default.aspx/TranshumanSpaceWiki/Infomorph.html

7. http://cci.mit.edu/about/index.html