Research and engineering focused on building and improving systems that perform tasks associated with intelligence, such as image recognition, language modeling, and control.

A subfield of philosophy of mind and ethics that examines conceptual and normative questions raised by AI, such as consciousness, understanding, moral status, and responsibility.

The subjective, qualitative aspect of mind; there is something it is like to be a conscious system, to have experiences from a first-person point of view.

Functionalism (in philosophy of mind)

The view that mental states are defined by their functional roles (what they do and how they interact) rather than by the physical material they are made of.

A situation where it is unclear who should be held morally or legally responsible for the actions or harms involving AI systems.

A European Union regulation, politically agreed in 2023 and entering into force in 2024, that introduces a risk-based framework for governing AI systems.

What is Technical AI?

Computer science and engineering work aimed at building and improving systems that perform tasks associated with intelligence.

From Science Fiction to Serious Philosophy: Why AI Forces Us to Rethink Mind and Morality — Thinking Machines: Philosophy, Consciousness, and the Ethics of AI

Setting the Stage: Why AI Is Now a Philosophical Problem

AI Leaves the Lab

Since around 2018, AI systems have moved from labs and science fiction into everyday life: recommender systems, chatbots, image generators, and autonomous vehicles are now common.

Old Questions, New Pressure

These systems put pressure on old philosophical questions: What is a mind? What is consciousness? Who is morally responsible when AI systems cause harm?

Beyond Engineering

These are philosophical questions because they concern concepts, values, and justification, not just how to build or optimize systems.

Module Goals

You will learn to distinguish technical AI research from philosophy of AI, see how AI challenges ideas of mind and morality, and connect current technologies to these debates.

AI as a Test Case

Treat AI as a test case: if our theories of mind and morality cannot handle AI, those theories might need revision.

Technical AI vs Philosophy of AI

Two Different Conversations

Technical AI and philosophy of AI are related but distinct. One builds systems; the other asks what those systems mean for concepts like mind and morality.

Technical AI Focus

Technical AI asks: How do we build models that recognize images, answer questions, or drive cars more accurately, efficiently, and safely?

Philosophy of AI Focus

Philosophy of AI asks: Could an AI be conscious? What is understanding? Who is responsible when AI causes harm? Should AI systems have moral status?

Conceptual vs Normative

Conceptual questions clarify meanings of mind, person, and intelligence. Normative questions ask what we ought to do and how we should regulate or treat AI.

A Simple Contrast

Technical AI: "Can we build it, and how?" Philosophy of AI: "What is it, and what should we do about it?"

A Short History: From Mechanical Brains to ChatGPT-era AI

Early Mechanization

Automata and calculators showed that some reasoning could be mechanized. Leibniz even imagined a "reasoning calculus" to settle disputes by calculation.

Turing's Shift (1950)

Alan Turing proposed the Imitation Game (Turing Test), focusing on whether a machine's conversation is indistinguishable from a human's, not on abstract "thinking".

Symbolic AI (GOFAI)

From the 1950s to the 1980s, symbolic AI tried to capture intelligence with explicit symbols and rules, raising questions about whether this yields true understanding.

Neural Networks and Learning

From the 1980s, connectionism and machine learning emphasized pattern recognition in data, inviting comparisons to how brains might work.

Deep Learning to LLMs

Since about 2012, deep learning has dominated. From roughly 2018 onward, large language and multimodal models made AI socially visible and philosophically urgent.

Current AI Capabilities and Limits (2026 Snapshot)

Strength: Pattern Recognition

AI excels at pattern recognition: classifying images, detecting faces, and generating realistic pictures, such as flagging possible tumors in medical scans.

Strength: Language Generation

Large language models write fluent text, summarize, answer questions, and generate code, helping users draft essays or debug programs.

Strength: Games and Control

AI beats humans at Go and chess and can control cars on highways or robots in warehouses, especially in structured environments.

Limit: Hallucinations

LLMs can hallucinate: they output confident but false claims, because they predict text from data rather than consult a grounded world model.

Limit: No Consensus on Consciousness

There is no accepted evidence that current AI systems are conscious. Most researchers see them as powerful pattern machines, not experiencing minds.

Limit: Alignment and Control

AI can be biased or unsafe. Aligning behavior with human values and ensuring control is a major research and policy problem.

Mind and Consciousness: What Would It Take for AI to Have a Mind?

Intelligence vs Consciousness

Intelligence is problem-solving and learning. Consciousness is there being something it is like to be a system. An AI might be intelligent without being conscious.

Functionalism

Functionalists say mental states are defined by what they do, not what they are made of. On this view, a silicon system could have a mind if organized correctly.

Biological Views

Some theorists claim consciousness depends on biological features of brains, suggesting digital AI may never be conscious, regardless of complexity.

LLMs as Test Cases

Large language models imitate conversation but lack bodies and lived histories. Are their behaviors enough to count as having minds?

Current Consensus

As of 2026, most philosophers and scientists think existing AI systems are not conscious, though debate continues about what would be required.

Key Skill

Learn to distinguish what a system does from what, if anything, it is like to be that system. Do not equate fluent behavior with experience.

Moral Responsibility and Agency: When Machines Act, Who Is Responsible?

What Is Moral Agency?

Moral agency typically involves understanding moral reasons, acting on them, and some kind of awareness or consciousness.

Current AI and Agency

Present-day AI systems optimize objectives set by humans but do not understand moral reasons. They are not generally treated as moral agents.

Responsibility Gaps

When AI harms, developers, users, and companies may each deny full responsibility, creating a responsibility gap that law and ethics must address.

Legal Responses

The EU AI Act, entering into force in 2024, and global guidelines from OECD and UN stress human accountability and risk-based obligations.

Philosophical Takeaway

Despite AI autonomy, we still see humans as the main bearers of moral and legal responsibility for AI design, deployment, and oversight.

Thought Exercise: Chatbot, Tool, or Person?

7. Thought Exercise: Chatbot, Tool, or Person?

Work through this scenario step by step. You can pause after each question and write down your answers.

Scenario

A university deploys an advanced chatbot to answer student questions about coursework, mental health resources, and administrative procedures. Students often say things like, "It really understands me" and "I feel like I am talking to a friend."

One night, the chatbot gives a student dangerously misleading advice about handling a crisis. The student is harmed as a result.

Step 1: Classify the system

Ask yourself:

Is this chatbot more like:

A tool (like a calculator),
A service (like a call center), or
A quasi-person (like a very limited artificial agent)?

What features make you lean one way or another (language fluency, apparent empathy, lack of consciousness, etc.)?

Write 2–3 sentences justifying your classification.

Step 2: Responsibility analysis

Consider at least three actors:

The university administration.
The developers of the chatbot.
The student who used it.

For each, answer:

What could they reasonably have foreseen?
What steps could they have taken to reduce risk?
How much moral responsibility do you assign them (low, medium, high)? Why?

Step 3: Mind and morality link

Finally, connect this to mind and morality:

Does your judgment about responsibility depend on whether the chatbot has a mind or consciousness?
Would your view change if you became convinced that the chatbot genuinely understands the student and can suffer?

Write a short paragraph (4–6 sentences) explaining how your answers would change if the chatbot were conscious versus if it is just a pattern machine.

Check Understanding: Mind and Responsibility

8. Check Understanding: Mind and Responsibility

Answer this question to test your grasp of the distinction between technical AI and philosophy of AI.

Which of the following is MOST clearly a question in the philosophy of AI (rather than a purely technical AI question)?

How can we reduce the error rate of an image classifier on medical scans?
Could a large language model ever be genuinely conscious, or is it only simulating understanding?
What neural network architecture gives the best performance on a translation benchmark?
How can we compress a model to run efficiently on a smartphone?

Show Answer

Answer: B) Could a large language model ever be genuinely conscious, or is it only simulating understanding?

Option 2 is a philosophy of AI question: it asks whether a large language model could be genuinely conscious or only simulating understanding, which is a conceptual and philosophical issue. The other options are technical questions about performance and efficiency.

Key Term Review

9. Key Term Review

Use these flashcards to reinforce the core concepts from this module.

Technical AI: Research and engineering focused on building and improving systems that perform tasks associated with intelligence, such as image recognition, language modeling, and control.
Philosophy of AI: A subfield of philosophy of mind and ethics that examines conceptual and normative questions raised by AI, such as consciousness, understanding, moral status, and responsibility.
Consciousness: The subjective, qualitative aspect of mind; there is something it is like to be a conscious system, to have experiences from a first-person point of view.
Functionalism (in philosophy of mind): The view that mental states are defined by their functional roles (what they do and how they interact) rather than by the physical material they are made of.
Responsibility gap: A situation where it is unclear who should be held morally or legally responsible for the actions or harms involving AI systems.
Large Language Model (LLM): A type of AI system trained on massive text datasets to predict and generate text, capable of tasks like conversation, summarization, and code generation.
Moral agency: The capacity of an entity to be a subject of moral evaluation, typically requiring understanding of moral reasons and the ability to act on them.

Key Terms

EU AI Act: A European Union regulation, politically agreed in 2023 and entering into force in 2024, that introduces a risk-based framework for governing AI systems.
Moral agency: The status of being an agent who can be held morally responsible, usually requiring understanding and responding to moral reasons.
Technical AI: Computer science and engineering work aimed at building and improving systems that perform tasks associated with intelligence.
Consciousness: The subjective, experiential aspect of mind; what it is like to be an experiencing subject.
Functionalism: A theory in philosophy of mind that identifies mental states with their functional roles rather than their physical makeup.
Philosophy of AI: Philosophical study of conceptual and ethical questions raised by artificial intelligence, including mind, consciousness, and responsibility.
Responsibility gap: A gap that arises when harms involving AI systems occur but it is unclear who should be held responsible.
Large Language Model (LLM): A neural network trained on large text corpora to predict and generate text, used for conversation, summarization, and related tasks.