AI on Trial

Foreword
Acknowledgements
Introduction: AI at a tipping point
Part 1: Opening Submissions
Chapter 1 A question of definition
What is meant by the term ‘Artificial Intelligence?’
What is the nature of our relationship with machines?
How do we move beyond one-trick ponies?
A case of the blind leading the blind?
To be, or not to be, AI: that is the question
Chapter 2 The state of things
What is the state of AI at the moment?
How does it work and how are decisions taken?
How should we use AI and where should it be deployed?
Should there be limits to where AI can be used?
Can we trust AI at the moment?
Have we already lost control of AI development?
Imposing constraints of ensuring oversight?
So is the issue not just technology, but humans and humanity itself?
Chapter 3 The building blocks of responsible AI
Ethical standards
Biased AI
Unknowable AI
Inappropriate AI
The foundation for responsible AI
AI without context
Graphs as the fabric for context
Building robust AI with context
Situational flexibility
Building trustworthy AI with context
Fairness
Trust and explainability
Considerations for building responsible AI
Summary
Chapter 4 The ethics dimension
The ethical perspective
Does AI need regulation?
A question of transparency?
Do we need to be more humane?
Do we need to define intelligence in all of this?
Can ethical standards be hardwired into software?
Do we need more education and training in ethics?
Part 2: Evidence
Chapter 5 Owning the digital future: In the AI world to whom does your data belong?
The battle for our data
Capturing our digital DNA
Getting on top of the data
Regaining control of our data
The dangers of personal data control
The allure of whole body data
The data model of you
The life blood of the Metaverse
Who owns your metadata in the Metaverse?
The data at our core
A data trust: our data in our interest
The new world created by data
The new ‘data’ rights
Chapter 6 Patently obvious – the AI inventor: Should an AI be able to make IP?
The first intellectual victory of the machines
What is ‘not obvious’?
The incredible value of instantly actionable information
The Go game changer
What then conceptually is a skilled machine?
Replacement or enhancement?
Will it be in the interests of skilled people to use inventive machines?
The new inventive war of discovery
Data and technological muscle will extract the value from invention
An innovation to break a log jam or another addition to a morass?
Could patent AI transparency open up more inventive opportunity?
AI transparency should clean a system which is creating blocks to innovation
The legislative challenge of the changing innovation landscape
Chapter 7 AI and cyber security: Are we weaponising the internet?
An AI war we may find it hard to prevent
The threat posed by relying on systems we cannot understand
Can we teach the machines to be good?
Should certain systems be designated as sacrosanct?
Out of our hands and out of control?
The criminal AI threat
The automated cyber war
The terrifying threat of an automated conflict beyond human control
The essential legal log
Taking responsibility for arm’s length harm
Combating the myth of the god in the machine
Chapter 8 AI as the information weapon: The data war to end
The very personal dangers of data
The information war
AI as the new world influencer
Hacking humanity
Encrypt data to save our souls
In the twenty-first century, the power is with the processors
Hi-tech brutal and efficient war of our times
In an AI world conflict will be a literal battle of hearts and minds
The technology arms race we have to face
Securing our data to protect our future
The very real consequences of virtual attacks
The internet and AI are now mainstream
Chapter 9 Driving an ethical approach to AI coding: Do we need a Hippocratic Oath in AI?
Do no evil
Imposing higher ethical standards on machines than on humanity
How far should technology intrude into our lives?
Ethics by design
The legal personhood of the decision maker
The moral issue of decision-making in certain contexts
AI accountability
How far can AI intrude into our data?
Training an AI system to make judgements on people
Abrogation of responsibility to the machine
Should machines have a ‘rush of blood’ built into them?
The real danger of a worldwide human inferiority complex
Appendix
Chapter 10 In pursuit of diversity: Is AI racist?
The startling discovery: owing to data AI can be racist
The devil is in the detail
Should we clean the data and start again?
Who decides upon the filter?
Big tech’s masked intervention in presentation
Auditing the AI
A statement of Intent: what is the AI meant to do?
The controversial argument for more data
Transparency essential for understanding and public trust
Do we need an FDA for AI?
The need for multi-disciplinary oversight
Regulating AI holds our behaviour up to an ethical lens
Leaving bias in AI systems will be a legal time bomb
Going back to the classroom
Chapter 11 Decoding inherent unfairness: Can we correct hardwired bias?
AI targeting: us or them?
The missing names and faces
Poisoning the data well
The value of data
The background behind why data is collected
Predictive social control
The very real dangers of the metadata
Dealing with a bolted ‘data horse’
The dangers of automated accidental misinterpretation
A road testing culture that threatens a motorway pile-up
Time to say ‘no’ to the computer
AI in whose interests?
A question of data
The call for an AI moratorium
The subjective intent of AI systems
Chapter 12 Gaming the system: Is AI being used to gain competitive advantage?
Beating AI at its own game
Preventing system manipulation by unethical humanity
Economic algorithmic warfare in the public interest
An algorithm Tsar?
An International Nuclear Weapons Authority for AI?
Who or what decides which information is good or bad?
Exposing bias and intent
The information war
Do bad people make bad AI?
What should the qualification for judgement be?
Where should we use AI?
Are human ethics higher than those of an AI?
Is it AI or not?
Can AI achieve contextual judgement?
Can AI interpret scientific judgements?
Chapter 13 Out of our hands?: Does AI already run our lives?
AI holds more than the promise of intelligence
Fake news and online harms
Who is responsible for the learning of an algorithm?
I think; therefore I am
Digital evolution
Run by robots
Our surrender to the machines
Regulating for our survival
Damned if we do, damned if we don’t: we must have AI
A new occupationally therapeutic world
Chapter 14 To be or not to be: the rights and responsibilities of AI: To whom does AI answer?
Is it actually AI?
A fine balance of competing interests
Rules are not the answer, context is everything
The human touch
The beauty of bias
The importance of human transparency
Where to be and where not to be
AI – the new age of enlightenment and opportunity?
The challenge of data use
What price an AI decision?
Chapter 15 Putting our lives in the hands of technology: Can we trust AI?
Data, data everywhere but not knowing how AI thinks
No regulation without technological comprehension
When is data good and when is it bad?
Just what data do you need?
The thorny moral question of data relevance
An argument for augmented intelligence?
Should jobs be protected against progress?
Poor human decisions preferred to those of machines
Transparency over data gathering
Auditing the algorithm
The discrimination of data
The case for an AI regulator
Free speech and the autocracy of the algorithm
What should we trust in technology?
Part 3: Closing Submissions
Chapter 16 Civil liability for AI
The importance of definition
The nature of legal liability
The basis of establishing liability
Accountability
A defensive strategy?
Chapter 17 Time for an AI law?
Why regulate?
Robot babies
Robot carers
Robot relationships
Biological robots
The robot professional
Robot and AI worlds
The Metaverse
The smart city
Smart neighbours
The dark side of the smart world
Revealing patterns in data
Data dependency
Sentient AI
Part 4: Final Verdicts
The View from the Bench: Lord Sales, Justice of the UK Supreme Court
Manifesto for Responsible AI: A blueprint for an ethical, legal and regulatory framework
Index