Information

This article is written from the perspective of the interests of this web site and its notion of a historical period called Informatia. For a more general contextual account see the Wikipedia articles on information and artificial intelligence. 

We live in an exciting Information Age and ‘information’ is the philosophical flavor of our times, a complex concept appropriate for our period in history, referred to on this web site as Informatia. Most of us are aware of the gathering influence of information on our lives through the rapid recent global uptake of the internet, smartphones, and social media. But the world’s academic knowledge is now available to all through Wikipedia. Anyone interested in any subject or discipline and who has access to the worldwide web can read and learn.

As a new phenomenon of our times we need to understand not only the social implications of this sudden democratization of knowledge but what exactly we mean by ‘information’. Not surprisingly it appears that along with complex technology have come complex ideas. information is well entrenched in biology, and especially genetics where coding, translation, editing

What is information?

When I say to you ‘Pass the salt please’ or ‘dog‘ . . . the information that you receive and comprehend is not the sound vibrations in the air. And if I wrote these statements on a piece of paper, it would not be the words (the letters of the words written in ink) – or even the molecules out of which the ink was made, that constitute the ‘information’ that has been conveyed. And, sent electronically, it would not be the pixels on my computer screen.

A painting is not just the pigments out of which it is composed. DNA is not just organic molecules. Novels and poems are not just the ink on the page, they convey to us much more than this. Can we express this extra meaning by saying that it is ‘symbols, sounds, or objects arranged in a particular way‘?

In sum, information is not the symbols and materials that are used to transmit it: it is more like ‘meaning’ or ‘content’.

So, what is information?

It seems we cannot start our investigation of this topic without some guidance or definition as a starting point.

But the discussion surrounding ‘information’ is young, with a rich and fascinating diversity of views. So, instead of looking at these gradually and systematically, let’s just jump straight into the cauldron of ideas before sifting and sorting.

Constructor Theory

The study of information has become strongly associated with what is known as constructor theory.^[1] Information Is a form of explanation in fundamental physics that has been devised by Professor David Deutsch and developed by Chiara Marletto, physicists at Oxford University.

Conventional procedure in physics is to declare the initial conditions and states of a system and then describe what happens. Constructor theory begins by assuming what is, in principle, possible and impossible according to the laws of physics. What actually happens then becomes an emergent property of the possible.

Science is then formulated as a series of tasks performed by constructors as objects that have the retained capacity to perform tasks. The world is then described in terms of transformations (change) in which something is changed (the substrate) and there is something that changes it (the constructor). This is a new way of formulating physics that makes ‘tasks’ the focus of attention. Information (as a constructor) is, for example, the software directing a robot. Knowledge (as collective learning) is a powerful and preserved constructor, as is DNA.

Constructor theory has been found useful in the field of quantum computation. Natural selection acts as a constructor acted on by the environment.

Fundamentalism

‘If you are carrying a hammer then everything looks like a nail’

What place or role does information hold in the physical world? Is it just a tool or metaphor – simply a convenient explanatory device – or is the bit-flipping of a computational substrate part of the fundamental fabric of the universe? Though traditionally treated as an aspect of knowledge and meaning and therefore derivative, could it be primary stuff? We can, for instance, readily understand that organisms are information-processing systems.

The traditional view of physicists is that the world consists of matter, translated scientifically as energy. But are mass and energy simply forms of information as a simpler, more economic, and therefore more fundamental idea or entity? Could information tell us the form that energy is taking (what it does) as it comes directly from the system itself – is the use of information a step closer to a detached account of the ‘reality’ of the universe?

First, there is the desire for grounding which comes in many forms expressed, for example, as the desire for foundations, axioms, or laws. This  approach provides points of explanatory departure, as Aristotle stated. Parsimony and elegance then suggest that one overarching physical foundation or fundamental principle is to be preferred. This also leads the search for a unified theory of everything.

Secondly, the search for foundational principles and ingredients of the universe has followed the tendency within science to find solutions by analysis, by investigating the relations of parts within wholes. This has led to the location of foundations in small and simple things (smallism) – like fundamental or elementary particles, bits,  numbers and so on.

The position developed on this web site, called aspect theory, is that scientific fundamentalism is misguided, and for psychological reasons. It is part of our innate mental structuring of the world that we both classify and prioritize the objects of our cognition, giving greater or lesser weight to some objects over others (rank-value) to form hierarchies. This is necessary for our survival. This is, however, an explanatory (epistemological) device: the world itself  does not make such distinctions. Everything in the world (?reality) exists equally (a flat ontology). On this view, to say that the world consists ‘fundamentally’ of numbers or bits of information is as absurd as it sounds: it may be consistent with a particular explanatory frame (physics, mathematics and computation), but there are other frames. Does this lead to relativism – ‘my frame is as good as your frame?’ Yes, but not in a negative way. Any frame is only as good as its efficiency in achieving the purpose for which it was designed. Numbers and bits of information might be meaningful for mathematicians, physicists and those immersed in computation, but these are not a viable currency for explaining the structure and function of living organisms. But this summation is nevertheless just one explanatory ‘perspective’ or ‘aspect’ of ‘reality’: it is just one way of describing the world, albeit a very useful one.

If this view has merit then we need to look elsewhere to find additional values for the notion of information. Scientists might also benefit from a general education in the innate predispositions of the human mind.

Definition

The etymology relates to the formation of an idea which suggests that it is not only fundamental in some way but it also has form and meaning but it needs an agent because there can be no value to information unless it has meaning and there can be no meaning without an agent. There must be an encoder and a decoder (a sender and receiver, a transmitter and a target)and a meaning that depends on context. Information is thus a relational and subjective concept with the potential to be assigned meaning. It can distinguish two states. When meaning is the key factor then this may vary according to the agent. But a non-conscious cell can distinguish between acid and alkaline conditions. It tells us ‘about’ something. It is the resolution of uncertainty.

This leads to a definition of information as a process, pattern, or connection, ‘a perceived difference that can make a difference’. It is a collection of facts or data, and sometimes the communication and reception of data which is accompanied by an increase in knowledge.

It is extremely difficult to deny the existence of information but to concede is to recognise an abstract non-physical object of universal importance. Since ‘information is neither matter or energy’ it is a novel and curious scientific object with an existence as puzzling as number and time. And perhaps on a par with these basic constituents of the universe.

Information in biology is what something is used for – its function; for humans it is transmission or communication.

Geological strata carry information. Is this different from the information carried in biological organisms like the genes that are a record or memory of past environments?

Artificial Intelligence

The most obvious feature of artificial intelligence is that it emulates the mindless natural process of evolution, the algorithm of life. It recognizes patterns and subjects these patterns to a process that adapts and refines.

‘AI’ is a general-purpose term for machines that can simulate human-like intelligence and perform routine tasks (as software robots or ‘bots’) that serve as substitutes for human cognition. They are therefore a form of information processing that can now be more efficient than human information-processing. AI is applied in many fields with the following selection providing an impression of its current application:

Chatbots as text-based or voice-based bots designed to engage in conversation with users and provide information or support; virtual Assistants to assist users with tasks, answer questions, set reminders, manage schedules, and perform various actions based on voice commands or text input; customer service to handle common inquiries, troubleshoot issues, and provide solutions; social media bots that interact with users on social media platforms, such as Twitter bots etc.; recommendation bots that provide personalized recommendations for products, services, movies, music, or content based on user preferences and behavior; personal finance bots to help manage finances, track expenses, and provide financial advice; health and wellness bots that offer health-related advice, track fitness activities, and provide mental health support; educational bots that assist in learning and education by providing information, answering questions, or offering personalized learning paths; gaming bots for video games that control non-player characters (NPCs) or provide assistance to and challenge to players; news bots that aggregate and deliver news articles or summaries based on user preferences; language translation bots that translate text or speech between different languages; content creation bots that generate content, such as writing articles, creating artwork, or composing music; task automation bots that automate repetitive tasks and workflows to increase efficiency; and robotic process automation (RPA) bots that mimic human interactions with software systems to automate business processes.

Stories and myths related to animated and intelligent machines date back to antiquity but the possibility gathered momentum with the Alan Turing’s pioneering work on computation and the advent of digital computers in the 1960s.

Development

Crucial to AI there is the capacity to ‘learn’.

This was achieved in 1952 when Arthur Samuel created the first self-learning program capable of playing checkers.

The term ’artificial intelligence’ was coined in 1956 and was soon followed by intensive studies in computer science and programing that lead to the development of expert systems, natural language processing, and machine learning algorithms. As computing power increased, AI research flourished in the 1970s and 1980s with advances in robotics and the creation of intelligent agents. However, progress and interest waned in the late 1980s and early 1990s until the emergence of the internet and Big Data revitalized the field.

The 21st century has witnessed a renaissance in AI, fueled by advancements in the understanding of neural networks, deep learning, and data processing.

Possibilities

Increasing general efficiency by speeding up routine intellectual and mechanical tasks. This could translate into a rise of as much as 7% in global GDP (New Scientist July 2023 3449, p. 3). Popular outcomes include the processing of natural language through ChatBots; self-driving cars; healthcare; finance; and education; improving decision-making. Integration of AI with other devices such as the speed and power of quantum computing.

There are many advances in mathematics to be made with increasing computer power, and limitless possibilities in the use of computer-generated graphics and animation.

Many aspects of daily life, such as health, education, and finances, can be personalized in a way that was not possible before – one example being the use of speaking bots like Siri, Google etc. Applied advances have been made in the determination of protein folding, drug development, research into nuclear fusion, study of enzymes such as those that break down plastic waste.

Concerns

Potential loss of jobs to computer creativity (authors, actors). Fears of superintelligence taking over control of human institutions. The generation of political propaganda, commandeering of elections etc.
Clearly some form of collaborative regulation are desirable.

AI assembles available information and arguments efficiently. This both challenges current research frontiers and makes frontier findings increasingly significant as they are rapidly absorbed into the knowledge base of collective learning.

There is debate about the concentration of computing power and resources in public or private hands especially as big money moves out of universities into private companies like Microsoft, Facebook, etc. This facilitates freedom of research but complicates any process of regulation.

Cybersecurity is a concern as hacking, scams, and cyberattacks increase along with the capacity to spread misinformation and disinformation as arguments, images, and videos become more convincing and persuasive.

Biological Information

Sentience

Among the most challenging intellectual aspects of AI are questions about sentience, qualia, and the hard problem of consciousness. Will machines ever be able to ‘experience’ the world in a manner even remotely resembling human experience. And how close to human agency can machine agency become? Certainly, machines can be constructed to be goal-directed, with human-like ‘intentions’ and ‘drives’. But there is, perhaps, a distinction here between ‘intelligence’ and ‘sentience’.

The idea that our minds are simply a physical extension of our bodies and senses is known as ’embodied cognition’, and the idea that consciousness exists by degree is readily applied to the animal kingdom.

Expanding Role of AI

AI is now generating quality writing, music, code, art, video, pictures, and much more as it continuously learns, corrects, and improves.

The emergence of AI marks the culmination of a process of knowledge and communication democratization. This is a progression that has passed from verbal to written forms, then to printed and digital media. Knowledge once possessed by few can now be obtained by the many.

Historically, knowledge once held only by priests and shamans was later placed in the minds of educated professionals in specialized fields like law, medicine, and religion, before the range of academic topics proliferated in specialist institutions. Now information, processed into a desirable form by AI, is available to anyone with access to the internet.

While software exists to identify AI-generated text, there are also competitive tools aimed at concealing its origin. The quality of AI-generated content is often exceptionally high, especially when tailored to specific needs. This raises little concern, except in cases where the originality and accuracy of the work are contested.
Given that these resources are accessible on personal computers and mobile devices, questions arise about the value of information platforms like Wikipedia and PlantsPeoplePlanet. After all, much of the content can be generated in more concise, comprehensive, or engaging formats by AI.

Use in education

Authority for quotations, factual claims etc. is usually given by citing the source of the material, information, or claim. Generative AI content often resists this convention because it has no specific author, and it usually cannot be authentically reproduced or recovered. The usual citation and referencing systems have not, at present, amended their style guides with directions on how to cite AI sources.
However, AI accesses and integrates a wide range of information that it can present in a concise way. It is therefore a valuable source of information in its own right. Students may therefore be asked to declare which generative AI tools have been used, to what extent, and the specific prompts used.

AI’s Limitations

While AI excels at tasks such as analyzing extensive datasets, identifying patterns, and generating hypotheses, it may struggle with creating entirely novel content or making substantial revisions that require nuanced contextual understanding and critical thinking. Therefore, human editing remains crucial. AI’s strength lies in its ability to accelerate progress and innovation through tasks like modeling, simulation, automation, and summarization. However, true original research often demands human experimentation and insights.

Glossary of common AI terms

AI Agent: A software entity that acts autonomously to achieve specific goals or tasks, often interacting with the environment.
Algorithm: A set of step-by-step instructions or rules followed by AI systems to solve specific problems or achieve particular tasks.
Artificial Intelligence (AI): A branch of computer science focused on creating machines that can simulate human-like intelligence and perform tasks that typically require human cognition. AI teaches computer systems to perform human tasks. Currently, most AI software is based on weak AI, which uses machine learning and deep learning.
Artificial General Intelligence (AI): AI that can challenge or out-compete humans in a range of cognitive tasks.
Bias: In AI, bias refers to the systematic and unfair favoring or disfavoring of particular groups or individuals by an algorithm due to biased training data or model design.
Bot: a contraction of the expression ‘software robot’ as an automated tool that carries out repetitive and mundane tasks
Computer Vision: An AI technology that enables computers to interpret and analyze visual information from images or videos.
Deep Learning (DL): A specialized form of neural network-like machine learning that analyzes complex data and make sophisticated decisions. It analyzes different data types (images, text) with minimal human input iterating multiple training rounds to analyze complex data features and make predictions. For example, recognizing a particular physical object in different situations.
Feature Extraction: The process of identifying and selecting relevant features or patterns from raw data to be used as input for AI models.
Generative AI: Artificial intelligence that can create output that resembles human-created content – text, images, videos etc.
Hyperparameter: Parameters set before the training process that affect the model’s learning process, such as learning rate and number of hidden layers in a neural network.
Machine Learning (ML): A subset of AI that enables machines to learn from data and improve their performance over time without being explicitly programmed.Natural Language Processing (NLP): The AI capability that enables machines to understand, interpret, and generate human language.
Neural Network: A type of ML model inspired by the human brain’s neural structure, consisting of interconnected nodes (neurons) that process and transmit information.
Overfitting: A situation where an AI model performs exceptionally well on the training data but poorly on unseen data, indicating it has not generalized well.
Reinforcement Learning: A type of ML where an agent learns to make decisions by interacting with an environment and receiving feedback in the form of rewards or penalties.
Supervised Learning: A type of ML where the model is trained on labeled data, meaning it learns from input-output pairs provided during training.
Test Data: A separate dataset used to evaluate the performance and generalization ability of trained AI models.
Training Data: The dataset used to train AI models, typically consisting of examples with known outcomes or labels.
Transfer Learning: A technique where knowledge gained from training one AI model is transferred and utilized to boost the performance of another related model.
Unsupervised Learning: A type of ML where the model learns from unlabeled data without explicit input-output pairs, seeking patterns and relationships within the data.

First published on the internet – 1 March 2019

. . . 2 August 2023 – begin section on artificial intelligence

Periodic table indicating the origin of each element. Elements from carbon to sulfur emerge in small stars, elements beyond iron in large stars, elements heavier than iron arise in supernovae
Courtesy Wikimedia Commons – Cmglee Acc. 10 Sept. 2015