Artificial Intelligence Demystified

At last – a clear explanation of what separates dumb computers from intelligent machines

When pressed, most software engineers will admit that they do not really understand how artificial intelligence (AI) can be real intelligence.

For people who have spent their entire careers working with logic and known outcomes the idea that their programs can be used to build truly intelligent machines seems rather unsettling.

If experts like these are struggling then what hope for other groups - like journalists, industry executives, regulators and policy makers?

Meanwhile, the leading players - Google, IBM, Microsoft, Facebook and Apple are racing off into the distance.

We therefore see an opportunity to help fill a major gap between the consensus view of what AI is, and what AI actually is.

Using a new approach we have generated a range of major insights that have allowed us to answer the most important questions in the field of AI in a new way which is clearer, simpler and more compelling than before, for example:

  • How can a computer that is based on logical programs ever be genuinely intelligent?
  • What is the difference between a dumb computer and an intelligent machine?
  • Is the intelligence of a machine limited by the intelligence of those who created it?
  • Is machine intelligence different in nature to human intelligence?

Executive Summary5

A new way to think about artificial intelligence (AI)6

Key findings10

Finally: Here is the difference between a dumb computer and an intelligent machine11




Technology Evolution: From Transistor to Watson17

Level 1: Transistor18

Level 2: Logic gate21

Level 3: Adder24

Level 4: 8-bit Adder26

Level 5: Arithmetic Logic Unit (ALU)29

Level 6: Single computer30

von Neumann Computer architecture: hardware30

von Neumann Computer architecture: software31

Creation of the software industry33

Level 7: Networked computing36

Level 8: Narrow AI systems37

Example 1: Melanoma Detection42


Presence of intelligence behaviour42

Example 2: Customer Targeting43

Closing remarks46

Level 9: Networked AI systems47

AI Web: A network of AI systems48

Collective AI (CAI)48


A definition of machine intelligence that works55

The Turing Test is insufficient to detect the presence of machine intelligence56

Searle's Chinese Room successfully invalidates the Turing  Test57

Algorithms cannot themselves lead to intelligence57

Regression cannot lead to intelligence, either59

But machine learning algorithms can exhibit intelligence61

Machine intelligence is fundamentally different in nature to human intelligence62

Quantum computing will lead to a different type of machine intelligence to that baseD on the transistor63

Machine intelligence is not either ‘present’ or ‘absent’64

Machine intelligence is as real as human intelligence and is growing exponentially65

The emergence of the AI web will be an inflexion point67

The strength of artificial intelligence exists on a sliding scale from ‘zero’ to ‘above human’68

Finally, what is the difference between a dumb computer and an intelligent machine?70


Dr Dave Taylor71

Andrew Sheehy71

Report Highlights

  • A new way to understand what AI is
  • Clear explanations that work
  • How AI arose
  • A new definition for machine intelligence (AI)
  • Explains term like weak AI, strong AI, AGI in a new way
  • The difference between machine and human intelligence.

Ordering Information

Title: Artificial Intelligence Demystified
Pages: 70
Updated: 16 Jun 2016
License: Single User
Format: PDF
Delivery: Email and Online.
Price: £295
Buy Now