Artificial-Intelligence-glossary

Artificial Intelligence (AI)

(Some incomplete notes. Read an article on AI in education)

Our provisional definition...

A system designed to emulate (the outward appearance of) human learning & problem-solving.

Some see AI as a co-creative tool, given that it can aggregate, draw inferences and reconfabulate the opinions of many speakers. But whereas human co-creativity harnesses the gut feelings of those facing a current predicament, AI machines only draw upon large bodies of written data from the past. In the Menexenus, Plato's idea of 'voices of the dead' may remind us that, while inert data may augment human wisdom it cannot replace it. In short, AI is allopoietic, rather than autopoietic. Here, society may need to decide whether to prioritise the co-creative skills of the living, or to continue training a new species of machine to outperform humans.

The big picture

Our species faces climatic, military and ecological threats to the livability of planet Earth.
It would be reasonable to expect universities to re-design themselves for their deep, long term purpose.

This context

This page offers a commentary on trends within the education system.
It is not intended to offer a comprehensive exploration of AI's potential in fields, such as:
- medicine, software coding or healthcare.
- virtual assistants like Siri and Alexa.
- personalized online content recommendations.
- fraud detection systems in banking.
- navigation apps using real-time traffic data.
On the other hand, we know our global supply chains depend heavily on digital, electronic networks and technologies that are prone to attack and failure.
Societal resilience will diminish if our systems of learning fail to think beyond our technological dependency.

A pedagogic arms race?

ChatGPT versus Gradescope

Currently, our education system reflects the short-termism of current economic and political orthodoxies.
Many universities appear to expect undetectable plagiarism as a fact of life
This has led to an AI arms race between teachers and learners.

How AI is being adopted in the UK

The Russell Group of Universities has approved some uses of AI (see comment) despite the fact that it encourages plagiarism.
- e.g. Anglia Ruskin University uses AI-Assisted Grading and Feedback and claim a 40% increase in what they call "grading efficiency"
The University of Suffolk offers AI adaptive learning systems for tailoring course material for individual learners.
The Open University uses AI to identify students who do not fulfil their expectation (they also claim this gives them a 40% improvement for academic achievement).
The University of Brighton uses Snippet-Based AI Learning to boost student learning retention by 38%.
The University of Hertfordshire uses AI in analysing skill gaps and personalizing learning strategies - they claim a 35% improvement in overall academic performance.
Buckinghamshire New University uses AI and claims that it has reduced plagiarism by 45%

Can AI support Real Intelligence

How useful is it to spare learners the task of looking for answers to questions that are unique to themselves?
Many applications are sophisticated augmentations of digital search techniques.
They obviate the otherwise troublesome routines of looking for known answers to verbally posed questions.
- (n.b. the Self Evaluating Learning Framework was designed to obviate plagiarism (Wood, 1992 & 2005)
- See my IDEAbase system (Taylor & Wood, 1997) that sought to enhance the creative agency of the user, rather than to offer the more mechanical (and parasitic) search faculties.

Limitations of AI in learning

Some research shows that handwriting aids some learning when compared with keyboard usage.
AI systems tend to offer answers that were previously deemed by humans to be useful, correct, optimal or average.
- This often means they may replicate the social or psychological biases of those used to train them.
An important aspect of AI's 'artificiality' is its allopoietic nature.
- The autopoietic nature of living systems mean they have no purpose, apart from surviving.
- In this regard, AI systems use algorithms to fulfil a specific pre-ordained purpose.

Disembodied Intelligence

Digital computer systems have no wetware (e.g. self-coordinating metabolic processes such as the endocrine system).
They are based on written codes that apply rules and meta-rules.
By contrast, humans not only think with their minds but, also, with their bodies.
- e.g. tacit knowledge is an aspect of human intelligence that informs how we reason.
- See also experiential, emotional and playful aspects of learning and knowing.

Codes and Algorithms

Loosely speaking, the AI project evolved from code-based bureaucracy, invented some 5k years ago.
- e.g. alphabetical writing and numerical accounting.
- see a critique of the culture of unit-based currencies.
An important characteristic of Living (Non-Artificial) Intelligence is its heuristic nature
- Some argue that it transcends algorithms (although see SelfAwarePatterns, 2015).

Voices from the Dead

Plato criticised alphabetical writing for seeming intelligent but repeating itself when challenged.
AI also depends on information or knowledge largely gleaned from the past.
Of course, computer-based writing can make texts appear to adapt to a new context.
Nonetheless, it is the same illusion but applied at an organisational level.
- e.g. The Turing Test is often cited to celebrate how good we are at simulating a living presence.
- However, this may mask the strong human tendency to anthropomorphise non-living things.
- e.g. Weizenbaum's Doctor program was based on the radically patient-centred dialogues of Carl Rogers (we are easily fooled).
- Phaedrus: Thamus says to Theuth: ". . . you, who are the father of letters, have been led by your affection to ascribe to them a power the opposite of that which they really possess. For this invention will produce forgetfulness in the minds of those who learn to use it, because they will not practise their memory. Their trust in writing, produced by external characters which are not part of themselves will discourage the use of their own memory within them. You have invented an elixir not of memory but of reminding; and you offer your pupils the appearance of wisdom, not true wisdom, for they will read many things without instruction and will therefore seem to know many things, when they are for the most part ignorant and hard to get along with, since they are not wise, but only appear wise." Quoted in Yates (1966) “The Art of Memory”