Roger Clarke's Web-Site
© Xamax Consultancy Pty Ltd, 1995-2026
Infrastructure
& Privacy
Matilda
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 | ||||
Version of 26 January 2026
© Xamax Consultancy Pty Ltd, 2025
Available under an AEShareNet 
licence or a Creative
Commons 
licence.
This document is at http://rogerclarke.com/EC/AIOP.html
Public concern about Artificial Intelligence (AI) and associated technologies is very high. This Opinion Piece summarises the reasons for those concerns and then argues the case for a reconception of AI, as an antidote to the foment they are causing. There have been serious problems with the notion of AI from its very beginnings, and these have compounded and deepened over the decades. The infectious enthusiasm of earlier times has been replaced by irresponsible misrepresentation of technologies and their capabilities. This document brings a sceptical eye to the subject-matter, as an antidote to the widespread excesses of over-optimistic providers and promoters, and the blind faith of a large proportion of users and user-organisations.
This Opinion Piece commences with a review of the state of information technologies during the first quarter of the 21st century. It then introduces the often vague notion of AI, identifies exemplars, discusses the kinds of languages used to develop them, and outlines the threats inherent in AI. The nature of the data on which the technologies depend is considered. Attention is then turned beyond intelligence and software to associated technologies in the areas of data analytics, content creation, and robots and cyborgs acting in the real world.
Finally, an alternative conception of the overall field is proposed, which enables an integrated view of the whole. This lays a foundation for sober assessments of how humankind can harness the ideas that are valuable, can identify and avoid ideas that are inappropriate and dangerous, and can manage the predictable harm and the contingent risks that are inherent in these powerful information technologies.
Recent forms of AI have unleashed intense public wonderment, fascination and fear. Putting aside the more fanciful prophecies of doom, there are strong grounds for concern that people are in the process of being swamped by uncontrolled technological change. Advanced information technologies (IT) do indeed both promise and threaten further substantial changes in economies, societies and polities. Those technologies also directly impact individuals' experience of, and interactions with, the physical world, and in and with virtual worlds.
My purpose is to outline the technologies and the mutually reinforcing nature of the interactions among them, and present them in a manner reasonably accessible to readers with only a moderate knowledge of the topic. Using that as a basis, I propose re-conception of the field in a way that redirects our energies towards constructive synergy between people and their tools. Links are provided to works that address specific topics at greater depth but in a accessible manner.
This Opinion Piece is intended to be accessible well beyond the computing and information systems disciplines and professions. As a starting-point, it therefore begins with a brisk review of the information infrastructure and tools of the early decades of the 21st century.
Key elements from the public's perspective are computer-based devices designed for use by an individual human. Broadly speaking, three categories are usefully differentiated:
Large-scale computing resources used by organisations are referred to as 'servers' and 'server farms', to distinguish them from individual users' 'client devices'. Increasingly, computing capabilities have also been embedded into other 'things' as diverse as motor vehicles, many components of buildings, free-standing cameras. and consumer appliances.
Support for communications capabilities among devices became entrenched between the mid-1980s and late 1990s, including 'wired' forms using metal and optical fibre cabling, and 'unwired' forms using various parts of the electromagnetic spectrum. Some channels are designed for communications within limited areas and others for tele-communications over distance. Some communications facilities are intended for general usage, and others for particular purposes, such as those reserved for emergency services.
A further, often overlooked function of hardware is the capability to directly impact the real world. The generic term 'actuator' refers to a component of an artefact that performs actions. It encompasses robotic arms, card-sliders in automated bank-teller machines (ATMs), the automated operation of sluice-gates in dams, programmed changes in traffic-lights, and the hydraulics of aircraft controls. The term is also applicable to workaday functions such as the automated despatch of communications (invoices, letters, emails).
Data is any symbol, sign or measure that is in a form accessible to a person or an artefact. The term 'empirical data' refers to data that is intended to represent some aspect of a real world phenomenon. For many years, data was processed and stored in what is currently referred to as 'analogue' forms, such as text and numbers written with pen on paper, and sound recorded on vinyl. The original meaning of 'analogue' was 'sampled continuously from the environment', whereas 'digital' implies sampling at time-intervals, producing a series of discrete data-values. However, the meanings have drifted, so that 'digital' now effectively means 'conveniently represented in binary form, and hence efficient to process and transmit', and 'analogue' effectively means 'not digital'.
The term 'data capture' has long been used to refer to expression in machine-readable form. A more appropriate descriptor in many cases would be 'data creation', because that better reflects the fairly common situation in which only a tenuous link exists between the real-world phenomenon and the data that is created to represent it. During the late 20th century, there was a focus on 'digitisation', encompassing the expression of new data in convenient, digital form ('born digital'), and the conversion of existing, analogue data into digital format (e.g. by scanning text into a digital image, and then using a optical character recognition tool to recover the text and store the text in digital form). The digital era since then has seen dramatically increasing volumes of data available for processing and transmission.
Some digital data is created by organisations' employees. Some is created automatically by computer-based systems, by combining data that those systems already hold with data arising from new transactions. Some data is also generated by 'analogue to digital converters' (ADCs), which detect a feature of the real world (such as temperature, movement, sound, light, or some other segment of the electromagnetic spectrum) and generate digital data to represent the phenomenon. Modern microphones use ADCs, as do digital displays of temperature and vehicle-speed.
Since the mid-to-late 1990s, the Web has been harnessed by organisations to transfer a considerable proportion of data capture effort and cost out to people external to the organisation that acquires it. Some of the captured data arises from conscious acts by customers, clients and other individuals using self-service facilities. However, people's actions generate a great deal of data without their knowledge. Data that can be associated with individuals has come to be transmitted and widely used by organisations without the consent of the people it relates to.
Prior to the period 1990-2000, most data-capture was performed for a specific organisation, and for specific purposes. The definitions of what each data-item meant, and how carefully its quality was assured, had little regard for any other purposes, and was in any case constrained by the data's value as perceived by that organisation. However, the quest for organisational efficiencies led to a desire for tight integration across organisational functions, and hence these 'silos' were broken down and data-collections were consolidated. Data was now to be managed as an organisational asset, irrespective of its origins, and of the interests of stakeholders other than the organisation that stored and processed it. Many organisations have paid limited attention to the data definition and data quality issues inherent in consolidated data-collections, which they, rather hopefully, regard as the 'single source of truth'.
The resulting, vast volumes of data that are now available are subject to a great deal of expropriation by additional organisations, and re-purposing to many and varied uses. Government agencies pass data among themselves, and corporations share and trade data with their 'business partners'. This breaks the linkage of data's meaning and quality with the original purposes of collection. Expropriated data from multiple sources is then subjected to data consolidation and inter-relationship with other data, with little regard for compatibility among the data-items' varying meanings and varying quality-standards. This results in a great deal of scope for misinterpretation, and for material errors in inferences drawn, decisions made, and action taken. Organisations frequently fail to ensure that expropriated, consolidated and re-purposed data has adequate reliability, quality and compatibility for the additional uses made of it.
Computing devices, actuators , communications links and data represent infrastructure, with latent power. Software utilises that infrastructure in order to exercise that power. Software is a generic term for individual programs and suites of programs that contain, or give rise to, precise instructions in the particular 'machine-language' or 'instruction-set' at the heart of each computing device. It is useful to distinguish three categories of software.
Firstly, the term 'systems software' refers to software deep inside computers, extending beyond the operating system (OS) to many further elements. It overlays the bare device and networks, and provides services to the other two categories of software. Some kinds of systems software are of direct value to users, including tools to manage documents and other files, such as Finder under macOS and File Explorer on Microsoft Windows.
The second category is 'software development software'. This includes compilers and interpreters to enable programmers to write software in a convenient language and have it converted into the relevant device's machine-language. Other important examples are editors for capturing and editing programs, and tools for testing programs and identifying the errors that are causing software to deliver results that are not consistent with the software designer's intentions.
It is the third category, 'application software', that users of devices are most familiar with. This applies the device's capabilities to humanly-useful functions such as managing a photo-album, a family-tree or financial matters, supporting message interchange with other people, enabling the preparation and editing of documents, conducting social and economic transactions, and accessing content. During the period 1975-2000, increasing numbers of people had access to personal computing devices that ran software under their control, and maintained their data locally. As will be discussed shortly, that scene has since changed.
The means whereby developers write software are usefully categorised into different language generations. The sequence below is primarily based on the level of abstraction away from the underlying computing device, but it also roughly corresponds with the chronological sequence of adoption:
All of these generations of software development tools are in use, with an ongoing overall tendency away from the earlier toward the later, more abstract approaches.
During the early decades of consumer uses of computing facilities, people used software and data on devices that they possessed and controlled. Since the beginning of the 21st century, however, IT providers have changed the architecture of computing and communications technologies. The industry refers to the new pattern using such terms as 'application software as a service', whereas from the user's perspective it can be described as 'remote software and data services'.
A number of key developments enabled IT providers to attract users away from locally-installed application software and locally-stored data, and to thereby gain ascendancy over consumers:
Through these mechanisms, IT providers have achieved customer-capture, withdrawing access to application software that runs on the person's own devices, and attracting people away from general-purpose computing devices towards provider-controlled appliances. Their purpose was to trap people into reliance on remote services such as Microsoft 365, Google Docs, digital music streaming, ancestry.com and most communications tools and social media; and they've succeeded.
This information infrastructure provides the basis on which a number of advanced, powerful and inter-linked technologies are intervening into, and even controlling, people's lives. Together, these technologies are promising and threatening changes of a highly transformational nature, disrupting and undermining existing cultures, societies and polities.
The term 'artificial intelligence' (AI) was coined 70 years ago. The meanings ascribed to the term have been many and varied. My contention is that, even in its most coherent form, it embodies, and always has embodied, a serious and harmful misconception. A major shift in the discussion is long overdue. This section begins the discussion by outlining three interpretations of what AI means. It then identifies some key exemplars of AI, and some tools used in the development of AI software. Finally, it offers an initial broad scan of the dangers inherent in AI.
In 1955, a group of researchers applied for funding for a symposium to discuss the use of a computing device to create artificial forms of human intelligence. The event stimulated many people's imaginations, and the term quickly entered the mainstream. Within five years, however, what started as 'conjecture' and 'hypothesis' had degraded into wild predictions. Enthusiasts now talked about duplicating the problem-solving and information-handling capabilities of the brain, and substituting machines for any and all human functions in organisations. This they confidently predicted would be achieved within one-to-two decades, i.e. by 1980. This original conception of AI can be seen as a 'grand challenge'; but to what purpose? There were close to 3 billion human intelligences on earth in 1955, and there are over 8 billion 70 years later. What benefit does humankind or the biosphere gain from devices that replicate all that human intelligence?
The original idea of replicating human intelligence has come to be referred to as 'artificial general intelligence' or 'strong AI'. Most AI practitioners have long since divorced themselves from that objective. They commonly distinguish their more practical approaches by referring to the original notion as having "aspired to replicate" human intelligence, whereas they regard themselves as being "inspired by" human intelligence.
AI practitioners tend to focus on particular ways of tackling particular categories of problem, and only occasionally venture into meta-questions about the nature of their endeavours. As a result, a sufficiently comprehensive and coherent definition of AI in contemporary practice is difficult to find. Based on multiple sources, here is this author's interpretation of what AI practitioners consider makes an intelligent handheld, robot, car or refrigerator :
Whether or not an application is AI software does not depend on which generation of software is used in its development. Among practitioners, the key factors that define AI are partly apparent world-awareness (1. "evidences perception and cognition") and partly embodiment in the software of apparent intentionality (2. "has goals" and 3. "formulates actions" to achieve them, and possibly 4. "implements those actions").
A third interpretation of AI appears to be achieving mainstream adoption among inter-governmental agencies. After some simplification of bureaucratic expression, the European Commission, the Council of Europe and the OECD consider an "AI system" to be a computer-based set of interacting processes that has some level of autonomy and some sense of objectives, and that draws inference(s) by processing various input data to generate output data. The manner in which computer-based inferencing is performed is different from human inferencing; but the outputs of both are applied to decision-making and action.
In comparison with this author's interpretation of AI practice, above, the bureaucratic definition adopts a weak form of 2. (in that "has some sense of objectives" is less than "has goals"), stipulates only a very limited form of 3. ("draws inferences" is less than "formulates actions [to achieve goals]"), and requires only "some level of autonomy", which only possibly extends to 4. ("implements those actions"). Only by vague implication does it encompass 1. ("evidences perception and cognition").
The variations among interpretations of AI are considerable. This ensures mutual unintelligibility of discussions about AI among communities that have diverse disciplinary and professional backgrounds, and diverse objectives. The final section of this Opinion Piece concludes that confusion will reign, technologies will run out of control, and serious harm will increasingly arise, unless and until two steps are taken: the abandonment of the misconceived notion of 'artificial' intelligence, and the substitution of terminology and definitions that reflect the human purposes to which the technologies are applied.
As a complement to that abstract and dry discussion of what AI means, it is helpful to identify some key exemplars of AI and AI systems that are in use, or that are being piloted. Many of the standout success stories from AI research have been in areas of pattern recognition. Progress in relation to images gave rise to optical character recognition (OCR), automated number plate recognition (ANPR), and computer vision. There have also been successes in the area of sound, such as music, by means of acoustic fingerprinting techniques.
The field of natural language processing (NLP), and in particular the rather ambitiously-named sub-field of natural language 'understanding' (NLU), analyses words and phrases in text, in order to extract a structured representation of the text's content. For example, it may recognise the nouns that define the subject-matter, and identify verbs and verb-forms that are related to the requestor's intent, thereby distinguishing a question (seeking information) from, say, an instruction to conduct a transaction. The end-result is a set of stored tokens that contains a model of the source-texts. That model is then available for further processing for various purposes. A rather different development has been natural language generation (NLG), whose function is to synthesise syntactically acceptable text from some structured representation of information.
During the 2020s, those two longstanding areas in the natural language space have been joined by a third. As a prelude to that third area, the field of machine learning (AI/ML) needs to be considered. This comprises a suite of techniques in which patterns and structures are postulated within a heap of data provided to them, and the patterns are expressed in an abstract model. That model is then used to classify new instances of data. The techniques are being applied to the drawing of inferences, which are then available to decision-making processes. Software development techniques for AI/ML are considered in the immediately following section, and the topic is revisited later, in the context of data analytics.
A recently-emerged category of AI applies similar techniques to AI/ML, but puts them to a different purpose. Rather than using models to categorise instances of data, as AI/ML techniques do, the purpose of Generative AI (GenAI) is to produce new instances of data that are intended to have particular attributes. GenAI models can deliver new instances of desired data intended to represent many different aspects of reality. For example, large volumes of audio (sound) data can be used to generate a model, which can be requested to synthesise new sounds. Those sounds might resemble (and perhaps be indistinguishable from) music, or a particular human's voice. Diagrams and animations can be produced in a similar manner. So can images, and hence video (which is merely a linked series of images, perhaps with an accompanying, synchronised sound-track). The currently very active area of GenAI models that handle text, such as the ChatGPT family of services, are considered in a later section.
As indicated earlier, AI software may be produced using any of the six generations of software development tool outlined above. In practice, many have been produced using 3rd-generation algorithmic or procedural languages. However, 5th-generation logic programming and rule-based expert systems have been convenient alternatives for many AI practitioners, and, since the late 2010s, 6th generation techniques have been intensively applied.
The 6th generation of tools is less driven by rational approaches to problem-definition, problem-solving and problem-domain modelling. Rather, the 6th generation exploits large volumes of empirical data by means that generally involve far less systemic reasoning about causality, and far more reliance on correlation. Scientific theories (those that deal in refutable propositions) embody systemic reasoning about causality. Software development up to the 5th generation reflects such theories. On the other hand, models that arise from applying mathematical or statistical processes to data have no semantics (in the sense of postulated relationships between data-items and real-world phenomena), and merely reflect correlations within a heap of data. So software development using 6th generation tools lack an underlying body of theory.
Multiple approaches fall within the 6th generation technique referred to as machine learning (AI/ML). The scale of computing power and high-speed memory reached a flex-point in about 2010. Since then, variants of artificial neural network (ANN) techniques have been the subject of a great deal of research, and active deployment. The following section considers their application to data analytics. The section after that discusses their use for Generative AI, in particular text-generating tools.
A particularly vital distinction among 3rd, 5th and 6th generation software development techniques is the extent to which the rationale underlying transaction outcomes is transparent or opaque. With 3rd-generation approaches, an algorithm or procedure exists, and from that a humanly-understandable explanation for an inference, decision or action can be readily constructed, e.g. 'After computing your remaining disposable income, and comparing it against the repayment schedule, it fell short of the threshold we require in order to provide you with that loan'. With AI/ML techniques like ANN, however, there is no rationale, and even if one existed at the time it could no longer be re-constructed.
In practice, only a small proportion of inferences, decisions and actions are disputed. Hence, in the large majority of cases, no-one calls for an explanation and none is offered. What matters, however, is that disputes are capable of being investigated. Although it took years for a test-case on the Australian government's 'Robodebt' scheme to come before a senior court, the software was developed using a 3rd generation tool, and so the illogic and hence illegality of the scheme could be exposed.
With a 5th generation tool such as those for rule-based expert systems, it may be theoretically feasible to incorporate into the design the generation of a humanly-understandable explanation. This might not identify all of the relevant factors, e.g. it might focus on a single disqualifier, such as age, without including mention of the fact that disposable income was also too low to justify a loan. (This is akin to the expression commonly seen in court judgments, along the lines of 'Having already determined the matter, I have no need to give further consideration to whether factors X, Y and Z apply').
The situation is very different in the case of the purely empirical approaches that are typical of 6th-generation development tools. No humanly-understandable rationale exists. Furthermore, such explanations as can be generated are of the nature of 'The state of the weightings on the nodes of the neural network at the time your data was run against it was such that your application was declined'. A field of 'eXplainable AI' (XAI) has arisen in an attempt to address this inadequacy. However, the field remains 'aspirational', meaning that little or no progress has been made to date.
The substantial absence of an ability to express a decision-rationale creates serious problems for a large proportion of applications produced using 5th and 6th generation software development techniques, especially contemporary approaches to AI/ML. Without a humanly-understandable explanation, it is infeasible for decisions to be reviewed and audited, and either justified or withdrawn. In short, accountability is undermined because affected parties, tribunals and courts cannot be provided with humanly-understandable explanations for decisions. This has substantial implications for the less powerful (typically, consumers, citizens and small business enterprises), but also for regulators, courts, policy agencies and parliaments, and for risk management and the insurance industry.
A great deal has been written about the threats that AI entails. Unfortunately, a great many different approaches have been adopted, and the opaqueness of many of the technologies invites confusion. The threats can be usefully clustered into five categories:
In summary:
AI commonly gives rise to errors of inference, of decision and of action, for which no rational explanations are available, and which may be incapable of investigation, correction and reparation.
The greater the degree of delegation of inference, decision and action to artefacts that embody AI, the greater the harm those errors inflict.
To address those threats, principles and business processes for responsible AI need to be established, and regulatory schemes for AI development and deployment need to be designed and implemented.
A wide variety of techniques have been developed to enable inferences to be drawn from data as a basis for decision-making and action. Such activities are variously referred to as 'data mining', 'data science' and 'data analytics'. For decades, much of the data that was available for analysis was in data collections created by a particular organisation for a particular purpose. The last few decades, however, have seen far more data created. Further, since about 2000, vast volumes of data have been traded or expropriated out of the data's original context. Proponents have waxed lyrical about the promise of such 'big data' for new insights, competitive advantage, and profit. Disparate data collections have been consolidated, and data analytics techniques have been applied to consolidated data heaps, with little regard for the diversity of data purposes, data meanings and data attributes.
In the 21st century, the challenges inherent in data analytics has been compounded by the increasing application of AI/ML. In one commonly-adopted approach, 'supervised learning', a (usually quite simple) model of some potentially relevant factors is provided to a piece of generic software. A 'training set' of data is fed in, resulting in weightings being associated with the various factors. The training-set may be carefully selected ('curated'), but it may be just a convenient sub-set of whatever data is available, or a big heap of data shovelled in indiscriminately. A populist version is along the lines of 'Feed the software enough pictures of cats and cat-like non-cats, with a tag on each saying whether it is or is not a cat'. The process of creating this artefact is referred to as 'machine learning', although 'computer-supported model development and adaptation' might be a better descriptor. New instances of data can be run against the resulting model, enabling each new instance to be categorised, and inferences drawn. A looser approach, referred to as 'unsupervised learning', relies heavily on the 6th-generation tool not just to compute weightings on predefined factors but also to postulate elements of the model. It represents something approaching 'automated model development and adaptation'.
Earlier techniques in data analytics were algorithmic or procedural in nature, and hence the rationale underlying inferences could be extracted and explained. In contrast, many of the most common AI/ML techniques that have been developed apply 6th-generation artificial neural networks (ANNs). As mentioned earlier, techniques of these kinds are not capable of generating humanly-understandable explanations. The combination of (a) the consolidation of data from disparate sources with diverse meanings and quality attributes with (b) opaque inferencing techniques, is a recipe for organisations to deceive themselves. This creates a high likelihood of an organisation making materially bad decisions, resulting in harm to themselves and to other affected parties, without the ability to investigate the sources of the errors..
Reference was made earlier to two relatively mature exemplars of AI: natural language understanding' (NLU) and natural language generation (NLG). During the 2020s, a new exemplar of AI burst into prominence, which combines with NLU and NLG to deliver new capabilities that are promising, exciting, superficially authoritative, and dangerous.
The term Generative AI (GenAI) refers to AI techniques that enable the production of synthetic content. The content might be in any form, including image, video and sound, or in 3D form such as holograms or additive manufacturing/3D printing. For data in textual form, successful synthesis has been based on a recently-emerged technique referred to as large language models (LLMs). An LM is simplistic in the sense that its approach to grammatical structure is limited to a means of, given a sequence of words, generating a likely next word. The apparent authority of the generated text derives from a combination of:
LLMs may shortly be complemented by, and perhaps ultimately replaced by, a generic syntax of the simplistic kind generated by LLMs, combined with less computationally intensive and hence less energy-wasteful small language models (SLMs).
Prominent text-oriented GenAI tools include ChatGPT, Microsoft's CoPilot front-end to the GPT LLM, Perplexity's interface to multiple LLMs, and the Chinese DeepSeek offering. Descriptions of GenAI for text are available in simplified form and in longer form.
The smoothness and apparent authenticity of responses provided by these tools lulls even experienced professionals into suspending their disbelief and disengaging their critical faculties. On the other hand, 20 potentially harmful attributes of GenAI have been identified, giving rise to many impacts, implications and risks. A sample of the raft of issues is as follows:
The evangelistic fervour of promoters, the willing suspension of disbelief by user organisations, and unreasoned, high levels of bandwagon investment are familiar preludes to a bubble-burst.
The interpretation of AI that reflects current practice includes a conditional fourth element relating to the software implementing actions that it has formulated. Computer-supported artefacts that embody actuators are commonly referred to as robots. The word 'robot' originated in Karel Capek's literary works at the end of World War I. Its usage since the 1940s implies both a machine with inherent computing capabilities, and a computer with sophisticated input/output devices. Isaac Asimov claims to have coined the term 'robotics' as early as 1940. He defined it in 1958, shortly after the notion of AI emerged, as "a science or art involving both artificial intelligence (to reason) and mechanical engineering (to perform physical acts suggested by reason)". From that base, he developed the insightful literary device that is Asimov's 'Laws of Robotics'.
The mainstream robotics industry of the early 21st century concerns itself with robots used in secondary industry, services and medical applications. It claims a worldwide installed base of about 4 million robots, over 60% of them in the automotive and electrical / electronic industry sectors. The majority are in China, Korea and Japan, and European adoption exceeds that of the Americas. Most installed robots operate primarily autonomously. A small percentage are designed as 'cobots', signifying collaboration with human beings. There are also providers of consumer drones, and of autonomous lawn-mowers and vacuum-cleaners, and a cyclical market exists for humanoid robots designed to physically and to some extent performatively resemble human beings.
Observation of successful applications of robots suggests a couple of guiding principles in relation to comparative robot advantage:
As indicated earlier, AI software can be implemented using any generation of software development tool. Robotics and AI have largely developed in parallel rather than intertwined, and it appears likely that the large majority of the installed base of industrial, services and medical robots does not use AI software at this stage, but rather software developed using 1st to 5th generation tools. On the other hand, AI is significant in computer vision, which is an important element of applications in transport and where robots need to handle objects whose location and orientation may vary.
With rapid growth occurring in drones and driverless cars, it is important to consider the nature of artefact autonomy. There is widespread distaste about the possibility of some kinds of decisions being made by artefacts. This exemplified by the much-discussed but insoluble category of moral dilemmas, sometimes relevant in transport contexts, that involve an unavoidable choice between two conflicting moral imperatives or harm to two values, such as the death of either one or more old people, or one or more children.
Meanwhile, other, generally low-level and real-time functions have already been delegated to software running in artefacts. In particular, vast numbers of autonomous actions are performed:
Importantly, the choice does not lie merely between control being exercised by either an autonomous machine or a self-sovereign human. There are varying levels or degrees of autonomy. An artefact can be designed to:
It is also possible to design a human-collaborative 'cobot' using decision support principles. In that case, the artefact, rather than taking action, provides recommendations, advises on options, or analyses options, and the human remains in control. In industrial robotics, for example, custom jobs such as repairs may need a human to cope with aspects of the work that are not within the device's specification.
There is a world of robotic behaviour beyond industrial, transportation and logistics applications. Many lesser machines have actuators installed, such as an ATM's sliding tray for chip-based debit and credit cards. Commercial and administrative applications can be programmed to make decisions, and to directly trigger actions by devices that have a direct impact on the real world. This may be as trivial as the despatch of an email, and the inclusion of an invoice or notice of a fine. Alternatively, the action may block a person's ability to make a payment, or to use public transport. Similarly, any person's mobility can be impeded through utilisation of features of contemporary motor vehicles, which are not only subject to continuous dataveillance but can also be immobilised remotely by third parties.
Affected parties, if they cannot discover a rational explanation for actions that have been taken, and if they can find no sentient being to deal with, face potentially serious difficulties, and worrying times. Organisations' control over, and accountability for, actions taken on their behalf is compromised. As robots proliferate outside controlled environments like production lines, dedicated bus-lanes and railway lines, and very-thinly-populated mining sites, effective regulatory schemes become a great deal more important.
Robotics involves creating an artefact that can act in the real world, and then providing it with some kind of processing capability, perhaps at a level or of a kind that can be thought of as intelligence. The inverse approach is for a human to be provided with some artefactual support, whether in relation to actions in the real world or of a mental nature. A prosthesisreplaces a human's physical capabilities, most commonly because those capabilities have been lost or materially reduced. A prosthesis may be separate from a human body (such as a walking-stick or a decompression chamber), or external but very adjacent (such as spectacles or a conventional hearing-aid). The term orthosis can be used for an artefact that augments rather then replaces a human's physical capabilities (such as a snorkel, SCUBA equipment, or a spacesuit that provides air but also protects against vacuum and radiation).
Some artefacts are capable of being implanted into a person's body, to replace lost or severely limited bodily function. Such an artefact is usefully described as an endo-prosthesis to replace missing functionality (e.g. a pacemaker, a cochlear implant hearing-aid, a retinal implant, or an implanted drug infusion device) or an endo-orthosis to enhance the human's functionality (a sprinter's artificial legs, or an infra-red detector implant). At this stage, implantation of artefacts that replace or enhance mental rather than physical capabilities appears to be still only emergent. An example would be implanted computer-aided visualisation, overlaying geographical contours over a view of a landscape, or displaying room-contents on the other side of an opaque wall. More ambitious 'neurotech' projects have been in train for decades, and arouse excitement from time to time.
The term 'cyborg' began in the 1960s with spacesuits, but quickly became captive to the entertainment industry. The term is usefully applied to a human being whose capabilities have been augmented by one or more technological artefacts that are more or less integrated with the physical person. Cyborgism is therefore more concerned with orthoses than prostheses, and particularly endo-orthoses. Like robotics, cyborgism is inherently linked to the fields of AI, action in the real and virtual worlds, and delegation of degrees of autonomy to artefacts.
In the Introduction to this Opinion Piece, it was contended that 'artificial intelligence' embodies, and always has embodied, a serious and harmful misconception, and that a major shift in the discussion is long overdue. The first aspect of the misconception was argued to be that pursuing a 'grand challenge' to create machine intelligence that was like human intelligence was a valueless endeavour, because there is no shortage of human intelligence. The need is not for copies of 'More of the Same' Intelligence, but for embodiment in artefacts of 'Usefully-Different' Intelligence.
A further problem is that the term 'artificial' carries with it unhelpful connotations of 'a copy of something real' / 'not as good as the real thing' / 'second-best'. A term such as 'Artefactual Intelligence' or 'Artefact Intelligence' avoids such overtones. It puts the focus on the purposes and outcomes of data processing performed by a manufactured thing.
Combining those ideas, 'Complementary Artefact Intelligence' has the following key attributes:
Ashby's 1956 ideas of intelligence amplification and Engelbart's 1962 proposals about augmenting human intellect have led to the term 'Augmented Intelligence'. This describes the combination of human intelligence with artefact intelligence that has been designed to be complementary to human intelligence.
The first line of the diagram below, read left to right, provides a graphical depiction of human intelligence blending with complementary artefact intelligence. That combination gives rise to a human-artefact synergy usefully described as augmented intelligence. The two notions of Complementary Artefact Intelligence and Augmented Intelligence provide a far more appropriate focal point for theory and practice than the hackneyed and ambiguous term 'artificial intelligence'.
These ideas also create the opportunity to link the three-part notion of augmented intelligence with the capability of physical actions. The second line of the depiction above shows human and artefactual means of affecting real-world things. The most commonly-used human effectors are fingers and thumbs. Artefacts' actuators were discussed earlier. Long ago, a human using their hand to grasp a hammerstone and strike it against a piece of flint gave rise to a far more capable Augmented Actor.
Rather than conceiving of a robot as a standalone artefact, it can be thought of as a repository of artefactual capability that is complementary to human physical capabilities. The third line of the depiction can be read as applying the cobotics notion. In addition to the original 'collaborative robotics', a further interpretation is 'Complementary Artefact Capability', which when combined with human capability delivers 'Augmented Capability'.
In each of the three lines, humans stay in control, and artefacts' physical but also intellectual capabilities are conceived as tools. The artefactual elements are exploited by humans in order to achieve more than humans could achieve without the artefacts. This approach applies to the 'AI'/Robotics space Ellul's 1964 ideas on technological society and Norman's 2000 ideas on 'The Design of Everyday Things'.
The information infrastructure of the mid-2020s is radically different from that at the turn of the century. The control that consumers had over their devices, their software and their data has been wrenched away from them, and into the hands of IT providers, such that:
Onto this provider-dominated infrastructure, further features are being overlaid. They include:
The public alarm about the encroachment of AI and associated technologies on humanity needs to be addressed in a constructive manner. The purpose of this Opinion Piece has been to provide an accessible analysis of AI and associated technologies, as a basis for arguing for the abandonment of the old notion of AI. That enables the substitution of a cluster of linked terms that can stimulate a more appropriate design ethos for information technologies and their application. That ethos needs to be reflected in each organisation's internal framework for the responsible creation and application of these technologies, in guidance published by regulatory or oversight agencies, and in law.
This Opinion draws on many sources. Rather than citing sources in-line, in the manner of a refereed article, and thereby interrupting the text flow, the approach adopted in this Piece is to provide within-text links that the reader can easily ignore (mostly) or click (when seeking more information). Many of the links are to predigested renditions of the particular topic previously published by the author (the large majority of which are refereed works including conventional citations). Where the author is aware of a better source that is accessible and authoritative, that source has been used. Where Wikipedia or other edited resource-pages have been used, they have been evaluated for reliability prior to being inserted.
To complement the within-text links, this Appendix lists the references used.
Ashby R. (1956) 'Design for an Intelligence-Amplifier' in Shannon C.E. & McCarthy J. (eds.) 'Automata Studies' Princeton University Press, 1956, pp. 215-234, at https://www.google.com.au/books/edition/Automata_Studies/A2CYDwAAQBAJ?hl=en&gbpv=1&dq=%27Automata+Studies%27&printsec=frontcover
Bender E.M., Gebru T., McMillan-Major A. & Shmitchell S. (2021) 'On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?' Proc. FAccT'21, March 3-10, 2021, Virtual Event, Canada, pp.610-623, at https://dl.acm.org/doi/pdf/10.1145/3442188.3445922
Brennen S. & Kreiss D. (2016) 'Digitalization and Digitization' International Encyclopedia of Communication Theory and Philosophy, October 2016, PrePrint at http://culturedigitally.org/2014/09/digitalization-and-digitization/
Clarke R. (1982) 'A Background to Program Generators for Commercial Applications' Australian Computer Journal 14,2 (May 1982) 48-55, at https://www.rogerclarke.com/SOS/PG-ACS-198205.pdf
Clarke R. (1988) 'Information Technology and Dataveillance' Commun. ACM 31,5 (May 1988) 498-512, PrePrint at http://rogerclarke.com/DV/CACM88.html
Clarke R. (1991) 'A Contingency Approach to the Software Generations' Database 22,3 (Summer 1991) 23-34, PrePrint at https://rogerclarke.com/SOS/SwareGenns.html
Clarke R. (1993) 'Asimov's Laws of Robotics: Implications for Information Technology -- Part I' IEEE Computer 26,12 (December 1993) 53-61 and 'Part II' IEEE Computer 27,1 (January 1994) 57-66, PrePrint at https://rogerclarke.com/SOS/Asimov.html
Clarke R. (1994) 'The Digital Persona and its Application to Data Surveillance' The Information Society 10,2 (June 1994) 77-92, PrePrint at http://rogerclarke.com/DV/DigPersona.html
Clarke R. (1999) 'Book Review of Shapiro & Varian (1999): 'Information Rules: A Strategic Guide to the Network Economy'' Xamax Consultancy Pty Ltd, 11 December 1999, at http://rogerclarke.com/EC/BR-IR.html
Clarke R. (2005a) 'Human-Artefact Hybridisation: Forms and Consequences' Invited Paper at the Ars Electronica 2005 Symposium on Hybrid - Living in Paradox, Linz, Austria, 2-3 September 2005, at https://rogerclarke.com/SOS/HAH0505.html
Clarke R. (2005b) 'Hybridity - Elements of a Theory' Xamax Consultancy Pty Ltd, May 2005, at http://rogerclarke.com/SOS/HAHTh0505.html
Clarke R. (2008) 'Web 2.0 as Syndication' Journal of Theoretical and Applied Electronic Commerce Research 3,2 (August 2008) 30-43, PrePrint at http://rogerclarke.com/EC/Web2C.html
Clarke R. (2011a) 'The Cloudy Future of Consumer Computing' Proc. 24th Bled eConference, June 2011, PrePrint at ../EC/CCC.htmlhttp://rogerclarke.com/EC/CCC.html
Clarke R. (2011b) 'Cyborg Rights' IEEE Technology and Society 30,3 (Fall 2011) 49-57, PrePrint at https://rogerclarke.com/SOS/CyRts-1102.html
Clarke R. (2013) 'The Shopping Cart Model' Xamax Consultancy Pty Ltd, September 2013, at http://rogerclarke.com/EC/ShoppingCart.html
Clarke R. (2014a) 'Understanding the Drone Epidemic' Computer Law & Security Review 30,3 (June 2014) 230-246, PrePrint at https://rogerclarke.com/SOS/Drones-E.html
Clarke R. (2014b) 'What Drones Inherit from Their Ancestors' Computer Law & Security Review 30,3 (June 2014) 247-262, PrePrint at https://rogerclarke.com/SOS/Drones-I.html
Clarke R. (2014c) 'Promise Unfulfilled: The Digital Persona Concept, Two Decades Later' Information Technology & People 27, 2 (Jun 2014) 182 - 207, PrePrint at http://rogerclarke.com/ID/DP12.html
Clarke R. (2016) 'Big Data, Big Risks' Information Systems Journal 26,1 (January 2016) 77-90, PrePrint at https://rogerclarke.com/EC/BDBR.html
Clarke R. (2017) 'Cyborgs Today' Xamax Consultancy Pty Ltd, October 2017, at http://rogerclarke.com/SOS/Cyborgs17.html
Clarke R. (2018) 'Guidelines for the Responsible Application of Data Analytics' Computer Law & Security Review 34,3 (May-Jun 2018) 467-476, PrePrint at https://rogerclarke.com/EC/GDA.html
Clarke R. (2019a) 'Risks Inherent in the Digital Surveillance Economy: A Research Agenda' Journal of Information Technology 34,1 (Mar 2019) 59-80, PrePrint at http://rogerclarke.com/EC/DSE.html
Clarke R. (2019b) 'Why the World Wants Controls over Artificial Intelligence' (1st in a series) Computer Law & Security Review 35,4 (2019) 423-433, PrePrint at https://rogerclarke.com/EC/AII.html
Clarke R. (2019c) 'Principles and Business Processes for Responsible AI' (2nd in a series) Computer Law & Security Review 35,4 (2019) 410-422, PrePrint at https://rogerclarke.com/EC/AIP.html
Clarke R. (2019d) 'Regulatory Alternatives for AI' (3rd in a series) Computer Law & Security Review 35,4 (2019) 398?409, PrePrint at https://rogerclarke.com/EC/AIR.html
Clarke R. (2020a) 'RegTech Opportunities in the Platform-Based Business Sector' Proc. Bled eConference, June 2020, pp. 79-106. PrePrint at http://rogerclarke.com/EC/RTFB.html
Clarke R. (2020b) 'Do Ethical Guidelines have a Role to Play in Relation to Data Analytics and AI/ML?' Proc. 9th Conf. Australasian Institute of Computer Ethics (AiCE 2020), Adelaide, Nov 2020, PrePrint at https://rogerclarke.com/EC/AIEG.html
Clarke R. (2022) 'Research Opportunities in the Regulatory Aspects of Electronic Markets' Electronic Markets 32, 1 (Jan-Mar 2022) 179-200, PrePrint at http://rogerclarke.com/EC/RAEM.html
Clarke R. (2023) 'The Re-Conception of AI: Beyond Artificial, and Beyond Intelligence' IEEE Trans. Techno. & Soc. 4,1 (March 2023) 24-33, PrePrint at http://rogerclarke.com/EC/AITS.html
Clarke R. (2025) 'Principles for the Responsible Application of Generative AI' Computer Law & Security Review 57 (May 2025) 106131, PrePrint at https://rogerclarke.com/EC/RGAI-C.html
Clarke R. (2026) 'An Evaluation of the EU AI Act against a Normative Framework for Regulatory Regimes' Working Paper, Xamax Consultancy Pty Ltd, January 2026, at https://rogerclarke.com/EC/RRE-AIA.html
Clarke R., Michael K. & Abbas R. (2024) 'Robodebt: A Socio-Technical Case Study of Public Sector Information Systems Failure', in Australasian J. of Infor. Syst. 28 (September 2024) 1-42, at https://journal.acs.org.au/index.php/ajis/article/view/4681/1481
Clarke R. & Wigan M.R. (2018) 'The Information Infrastructures of 1985 and of 2018: The Sociotechnical Context of Computer Law & Security' Computer Law and Security Review 30,4 (Jul-Aug 2017) 677-700, PrePrint at https://rogerclarke.com/II/IIC18.html
Ellul J. (1964) 'The Technological Society' Knopf, New York, 1964, at https://ia803209.us.archive.org/2/items/JacquesEllulTheTechnologicalSociety/Jacques%20Ellul%20-%20The%20Technological%20Society.pdf
Engelbart D.C. (1962) 'Augmenting Human Intellect: A Conceptual Framework' SRI Summary Report AFOSR-3223, Stanford Research Institute, October 1962, at https://dougengelbart.org/pubs/augment-3906.html
Norman D. (2000) 'The Design of Everyday Things' MT Press, 2000, summary at https://en.wikipedia.org/wiki/The_Design_of_Everyday_Things
Shapiro C. & Varian H.R. (1999) 'Information Rules: A Strategic Guide to the Network Economy' Harvard Business School Press, 1999
Zuboff S. (2015) 'Big other: surveillance capitalism and the prospects of an information civilization' Journal of Information Technology (2015) 30, 75-89, at https://cryptome.org/2015/07/big-other.pdf
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.
| Personalia |
Photographs Presentations Videos |
Access Statistics |
 |
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: 1 October 2025 - Last Amended: 26 January 2026 by Roger Clarke
This document is at www.rogerclarke.com/EC/AIOP.html
Mail to Webmaster - © Xamax Consultancy Pty Ltd, 1995-2026 - Privacy Policy