Description
Discover the truth behind AI's most dangerous myth: that machines can truly create
In The AI Illusion: Why Machines Aren't Creative, Luc Julia, co-creator of Siri and Chief Scientific Officer for the Renault Group, dismantles the hype surrounding generative AI by revealing what these technologies can actually do (as of today) versus what their promoters claim. Drawing on over 35 years' experience in the tech industry, Julia exposes the fundamental truth that generative AI doesn't create – it recombines existing data in response to prompts, producing impressive but ultimately derivative outputs that lack genuine creativity and understanding.
This essential guide takes readers on a comprehensive journey through AI's past, present, and future, systematically debunking seven pervasive myths that shape public perception of artificial intelligence. Julia examines the technical limitations, societal implications, and environmental costs of generative AI while providing practical insights into how these tools function and where they're headed.
The book:
- Reveals the technical reality behind generative AI's "hallucinations," biases, and inability to reason or understand language
- Exposes the environmental disaster created by energy-intensive AI training and deployment processes
- Analyzes the economic and employment impacts of AI adoption across industries and society
- Demonstrates why artificial general intelligence (AGI) remains scientifically impossible with current approaches
- Provides actionable solutions for more responsible AI development and regulation
Perfect for technology professionals, business leaders, policymakers, and curious readers trying to understand AI's true capabilities and limitations, The AI Illusion offers a clear-eyed perspective to help you navigate our AI-influenced future. It provides the critical thinking tools you'll need to see past the marketing hype and science fiction fantasies that dominate AI discourse.
Table of Contents
Preface xv
Part I The History of AI 1
1956: The Dartmouth Conference, Where It All Began 1
But Was AI Really Invented in 1956? 2
The Winter of AI: The First One 3
The Expert Systems: AI Is Back! 4
The First “Defeat” of Man Against AI 5
The Rise of Statistical AIs and Machine Learning 5
One of the First Image Recognizers in History 6
The Second “Defeat” of Man Against AI 8
Data: A Major Challenge 9
Tay: The AI That Went Awry 10
“Autonomous” Cars 12
What About GenAI? 16
GenAI for All of Us 18
The “True” AI 18
Part II Is Genai the Holy Grail of Technology? 21
A Revolution, but Not the One You Think 22
Good Choice of Words This Time 23
The Gartner Hype Cycle at Full Speed! 25
What Are the Concrete Applications for GenAI? 27
Weakness #1: Hallucinations 28
The One Prompt Too Many for Steven Schwartz 28
My Always Evolving Bio 29
Weakness #2: Lack of Accuracy 30
Does Being Wrong One Third of the Time Really Matter? 30
Weakness #3: They Can’t Think 31
Weakness #4: Jailbreaking— A Security Breach 33
What to Make of the Jailbreaking Story? 38
Solution #1: Fine-Tuning and RAG 39
Solution #2: Use the Data We Own and/or Trust 40
Is Data Theft Inherent on Training GenAI? 40
A Marketing Argument 43
Are AIs More and More Stupid? 44
Solution #3: Watermarking Is the Savior 45
Is It Possible to Watermark Text? 46
An AI to Control Another AI: Is It a Good Idea? 47
Solution #4: Open Source—A Source of Creativity 49
Solution #5: Small Language Models and Edge Computing 51
Solution #6: Hybrid AI and the End of GenAI 53
AGI Is Still an Inaccessible Dream 53
Part III the Seven Myths of Ai 55
Myth Number 0(Riginal) 56
AI Is Creative 56
What’s Creativity Anyway? 56
There Is “Create” and “create” 57
Can’t We Be Creative with an AI? 58
The Two Dimensions of Creativity 60
AI Will Never Innovate 61
Then AIs Are Useless? 62
Are AIs and Auto-Tune the Same? 63
Myth Number 1: AI Understands What It’s Telling You and Can Reflect Upon It 63
Cognitive Bias Plays Tricks on Us 64
The Characteristics of Intelligence 65
Is AI Really Better Than Humans at Recognizing Speech? 66
Are Humans Really Better Than AI at Understanding Language? 67
The Three Phases of Natural Language Understanding 67
AI Invents a New Language: Too Smart for a Human to Understand? 71
The True Failure of AI 73
Are AIs Better Than Humans at Communicating with Humans? 74
Myth Number 2: AIs Are Inexplicable Black Boxes 75
The Story of Gaston Julia and Fractals 76
The Unexpected and the Inexplicable: The Reason for the Black Box Myth 77
The Three Sources of the Unexpected 78
Example: The Screw-Driving Robot from Tesla 79
Is Explicability an Achievable Goal? 80
Myth Number 3: AI Is Going to Kill Us All 81
Terminator, by James Cameron 82
Avengers: Age of Ultron, by Joss Whedon 83
So, Killer AIs Do Not Exist? 84
Easter Eggs: The Poisoned Chalice of AI 85
The Story of the American Killer Drone 86
Scoop: Socrates Explains Why AIs Aren’t Yet Intelligent 87
GenAIs Need Humans 88
Myth Number 4: AI Is Objective 88
Color Blindness: An Example of a Subjective World 89
War: The Quintessence of Subjectivity 90
Biases Are Inherent in AI 90
AIs: Tools for Profit 91
The Ultimate Proof of AI’s Subjectivity 92
Myth Number 5: AI Will Lead to a Widespread Job Loss 93
There Have Been Other Revolutions Before AI 93
Innovation: Less Lethal Than We Think 94
Will Gen AIs Soon Be Executives’ Only Staff? 94
The Two Levels of AI’s Mastery 95
What Are the Consequences for the Organization of Companies? 96
AI: An Asset for Employability Rather Than a Hindrance 97
AI: The Least Job-Destroying Revolution 98
Myth Number 6: AI Can Learn Anything (Acquired versus Innate) 98
The Foundation of the Discussion 99
Descartes and the Beginning of Reconciliation 100
The Age of Enlightenment and the End of the Debate 100
And What About AI in All This? 101
Is Incomplete AI an Issue? 102
Myth Number 7: AI Cares and Does Everything Right 102
Ethics and Its Multiple Dimensions 102
Are GenAIs Built to Be Unethical? 103
Who’s Really in Charge of AI’s Ethics? 104
Is a Hammer Ethical? 105
Part IV What Ai Will Change in Our Lives 107
The Ecological Disaster of GenAI 108
The Three Energy-Hungry Activities 108
ChatGPT: Pandora’s Box of AI 112
Drying Up Humanity to Feed the Machine? 114
How to Meet AI Resource Demands 115
Data Centers: The Sine Qua Non Condition for GenAI 116
Some Shocking Figures on Data Centers 117
Greed Is a Bad Thing 118
There Is No Plan(et) B 120
AI Washing and the Gold Rush of GenAI 120
Getting Rich Thanks to AI or by Lying About It 121
AI Washing: Out of Greed or Fear? 122
The Tortoise and the Hare of GenAI 124
But What Does the AI Police Do? 126
“Out of Sight, Out of Mind” or Paying the Full Price 128
AI Forcing and the Obsolescence of Classic AI 130
The End of the GenAI Bubble? 132
Banning AI: A Necessary Evil? 133
Regulate, Yes; Ban, No 134
GDPR’s Shortcoming 135
What Does the AI Act Look Like in Detail? 137
Allow Technology but Ban Some Applications 140
Data Theft and AI 141
Who Is Really Responsible for AI-Generated Content? 143
AI and the Data Industry 145
Regulation’s Big Miss: The Ecology 147
Malicious Uses of GenAI 147
Fake News 148
What About Deepfakes? 149
Emotion Detectors 151
Hacking and Cybersecurity 152
What About GenAI? 153
A Quick Discussion on Quantum Computers 153
The Hidden Flaws of GenAI 154
Being Manipulated Without Realizing It 154
Cultures Cancelling 155
Evolution of Teaching and Working Methods 156
AI and Teaching: An Explosive Cocktail? 156
Is AI Going to Perform Tasks on Our Behalf? 157
An AI Monopoly? 157
Part V Our Future with Ai 159
Solution #1: Make AIs More Frugal 159
Solution #2: Regulate to Encourage Less Consumption 160
Solution #3: Limit Unnecessary Uses 161
AI and IoT 161
Conclusion 163
About the Author 165
Index 167
-
- 洋書電子書籍
- Robust Emotion Reco…
-
- 洋書電子書籍
- Engineering Statics
-
- 洋書電子書籍
- Sustainability in E…
-
- 洋書電子書籍
- Lost Souls : Manner…
