Roger Clarke's Web-Site © Xamax Consultancy Pty Ltd,  1995-2026
HOME	eBusiness	Information Infrastructure	Dataveillance & Privacy	Identity Matters	Other Topics
What's New		Waltzing Matilda	Advanced Site-Search

The Different Meanings of 'Artificial Intelligence'

Version of 27 July 2026

Roger Clarke **

© Xamax Consultancy Pty Ltd, 2025-26

This version supersedes the version of 24 January 2026.
This new version extends the introduction, and upgrades the 2nd half of the
previous version, splitting it off into a separate article on risk management

This document is at http://rogerclarke.com/EC/BAI.html

A great deal of enthusiasm for AI continues to be evident. But there are also plenty of warning signs. ChatGPT and its competitors (Generative AI / GenAI) stimulated claims that staff costs could be slashed. On the other hand, early studies suggest the claims may have been exaggerated, because higher-level skills are now in demand, to discover and fix errors in GenAI output. The massive amounts investors claim to be injecting have already given rise to talk of a stock market bubble, and predictions of another 'AI winter' are imminent. GenAI has been found to be prone to hallucinations, and humans may have been having some as well. Added to that, misunderstandings abound as to what AI is. It's a good time to take stock of the situation, and ensure that we, both as individuals, and in our organisational roles, are awake to realities, and understand and manage risks.

This document offers a brief introduction to what artificial intelligence (AI) means, addressed to the educated layperson. It lays the foundation for a separately-published proposal about how the (enormous) risks embodied in AI can be much better managed. The trick is to re-conceive the technologies, and put the focus instead on 'complementary artefactual intelligence', which melds with human intelligence to deliver 'augmented intelligence'.

Three different interpretations of 'artificial intelligence' interleave with one another, and confuse practioners, users, the media and the general public alike.

The motivation of AI pioneers, back in 1955, was to create artificial forms of human intelligence, hence Artificial Intelligence (AI). A few theorists persist with this approach, but most AI practitioners regard the notion of 'strong AI' as a wild, 'out there', even crackpot idea, and are embarrassed when entrepreneurs and marketers use that kind of language.

At the other extreme, contemporary intergovernmental bureaucrats in places like the European Commission, the Council of Europe and the OECD, refer to AI in such broad terms that it encompasses justabout all forms of software applications. Maybe we do need laws that say things like 'Organisations and individuals found to be reponsible for any software that kills people are liable for the harm done, and are subject to civil and criminal sanctions'. But if that's what regulators are saying, they can say it without confusing the meaning of AI.

The third interpretation of AI is that of practitioners in the art and science of AI. They do not aspire to replicate human intelligence, but rather are inspired by human intelligence in their quest for better ways to write software - and perhaps to architect and engineer artefacts that are even better than conventional computers at supporting effective software. Here's a definition that represents attitudes common among contemporary practitioners and theorists, particularly in the currently much-hyped and popular forms of data analytics using machine learning (AI/ML), and GenAI:

Definitions are one way to gain an understanding of a category of phenomena. Another approach is to look at examples, in the case of AI applications and of AI technologies and techniques.

Across its seven decades of existence, AI has had failures, modest successes and several big successes. From about 1970, 'logic programming' was developed, enabling statute law and government procedures to be expressed, and anomalies identified. That led, in the 1980s, to rule-based 'expert systems' as a convenient way to document and apply bodies of knowledge such as how to identify the particular bacteria causing blood infections. The broad area of 'pattern recognition' can be applied to images, sound and electromagnetic signals. Familiar forms of this kind of AI include the identification of printed characters in order to extract (optical character recognition, OCR) and the identifiers of motor vehicles (automated number/license plate recognition, ANPR/ALPR). There have also been successes in identifying pieces of music, by means of acoustic fingerprinting techniques.

Natural language 'understanding' (NLU) now copes well with the syntaxes of natural languages, although progress with semantics is much more challenging. Natural language generation (NLG) uses structured representations of information to synthesise text that is acceptable and even appealing. A technique referred to as 'artificial neural networks' (ANN) has been creating possibilities and challenges for the last decade. It has been applied to 'data analytics', which is the generation of new insights from existing data-collections. ANN has delivered surprises, some valuable, and some quite wrong. ANNs also underlie the current explosion in Generative AI, which produces highly convincing output, some of which after careful evaluation appears to be accurate, and some of which is dangerously misleading and even contains hallucinations.

Further Reading

For more detail on AI, see the more substantial backgrounder to AI and related fields. The current paper is a prelude to a second short paper on reconceiving AI in order to manage it. A fuller rendition of that argument is in an Opinion Piece, and the full argument is provided by a refereed paper in a semi-technical journal. An annotated index of papers by the author on AI topics provides access to underlying references numbering in the hundreds.

Author Affiliations

Roger Clarke is Principal of Xamax Consultancy Pty Ltd, Canberra. He is also a Visiting Professorial Fellow associated with UNSW Law & Justice, and a Visiting Professor in Computing in the College of Systems & Society at the Australian National University.

The content and infrastructure for these community service pages are provided by Roger Clarke through his consultancy company, Xamax. From the site's beginnings in August 1994 until February 2009, the infrastructure was provided by the Australian National University. During that time, the site accumulated close to 30 million hits. It had passed 80 million by the end of 2025. Sponsored by the Gallery, Bunhybee Grasslands, the extended Clarke Family, Knights of the Spatchcock and their drummer, and her Dr Nurse site

Xamax Consultancy Pty Ltd ACN: 002 360 456 78 Sidaway St, Chapman ACT 2611 AUSTRALIA Tel: +61 2 6288 6916

Created: 2 November 2025 - Last Amended: 27 July 2026 by Roger Clarke
This document is at www.rogerclarke.com/EC/BAI.html
Mail to Webmaster   -    © Xamax Consultancy Pty Ltd, 1995-2026   -    Privacy Policy