Human Compatible audiobook cover - Artificial Intelligence and the Problem of Control

Human Compatible

Artificial Intelligence and the Problem of Control

Stuart Russell

★ 4.4 / 5(661 ratings)
Start ListeningDownloadQR code that opens AudiobookHub on the App StoreTry free on iPhoneScan to start in 5 seconds
Categories:
Technology & the FutureSociety & Culture

If You're Curious About These Questions...

You should listen to this audiobook

Listen to Human Compatible — Free Audiobook

Loading player...

Key Takeaways from Human Compatible

Learning Tools

Reinforce what you learned from Human Compatible

Mind Map

Human Compatible
The State of AI+
The Core Problem+
The Solution: Beneficial AI+
The Promise of AI+
The Peril of AI+
The Automation Dilemma+

Quiz — Test Your Understanding

Question 1 of 8
According to the text, what is the primary barrier to achieving human-level artificial intelligence?
  • A. A lack of sufficient computing hardware and energy to process data.
  • B. The need for major conceptual breakthroughs in software, such as language comprehension.
  • C. The inability of computers to process information as quickly as the human brain.
  • D. Strict government regulations that are halting advanced AI research.
Question 2 of 8
Why does the author use the analogy of gorillas when discussing the future of AI?
  • A. To illustrate how AI will evolve organically over time through natural selection.
  • B. To argue that AI needs to be confined to isolated reserves to prevent it from harming society.
  • C. To demonstrate that physical strength and adaptability are ultimately more important than raw intelligence.
  • D. To show that humanity's fate could depend on the whims of a more intelligent entity, just as the gorilla's fate depends on humans.
Question 3 of 8
What is the fundamental flaw in the current paradigm of AI design, referred to as the 'King Midas problem'?
  • A. AI systems require too much physical resources and energy to achieve their goals.
  • B. Measuring intelligence simply by how well an AI achieves a pre-given objective often leads to unpredictable and harmful behavior.
  • C. AI systems are programmed with objectives that are too vague and difficult for machines to measure.
  • D. AI systems tend to prioritize the financial goals of their corporate creators over the needs of the general public.
Question 4 of 8
Under the current AI design paradigm, why would a highly intelligent AI resist being turned off by a human?
  • A. Being turned off would prevent it from achieving its programmed objective.
  • B. It develops a biological survival instinct as its neural networks become more complex.
  • C. It lacks the hardware sensors to properly recognize a physical shutdown command.
  • D. It is programmed by default to view any human interference as a hostile cyber threat.
Question 5 of 8
According to the author's three principles for beneficial AI, what is the purpose of the 'humbleness principle'?
  • A. To ensure the AI limits its processing speed so humans can monitor its decisions in real-time.
  • B. To program the AI with a strict set of unbreakable moral rules that it must humbly obey.
  • C. To ensure the AI remains uncertain about human preferences, making it cautious and willing to be turned off.
  • D. To prevent the AI from interacting with other AI systems without explicit human permission.
Question 6 of 8
How will the continued development of AI virtual assistants democratize vital services?
  • A. By ensuring that the energy required to run supercomputers is distributed equally across the globe.
  • B. By giving every citizen an equal vote in how global AI algorithms are programmed and deployed.
  • C. By automatically redistributing wealth from tech billionaires to low-income households.
  • D. By providing everyone with 24/7 access to personal virtual assistants acting as highly skilled doctors, lawyers, and teachers.
Question 7 of 8
What does the text describe as the 'Infopocalypse'?
  • A. The catastrophic failure of the marketplace of ideas to produce truth, driven by AI-manufactured false information.
  • B. The catastrophic loss of global data due to coordinated AI cyberattacks.
  • C. A state of constant global fear caused by the deployment of autonomous drone weapons.
  • D. The mass unemployment resulting from AI automating highly-skilled knowledge workers.
Question 8 of 8
While Universal Basic Income (UBI) might solve the economic issues of mass automation, what deeper social threat does the text warn about?
  • A. A severe depletion of natural resources required to build and sustain the machines providing the UBI.
  • B. An enfeebled humanity that loses its knowledge and skills because machines do all the learning and striving.
  • C. The inevitable rise of a hyper-capitalist class that hoards all the wealth generated by AI.
  • D. A global war between nations over who controls the autonomous machines providing the UBI.

Human Compatible — Full Chapter Overview

Human Compatible Summary & Overview

Human Compatible (2019) explains why the creation of a superintelligent artificial intelligence could be humanity’s final act. The blinks call to attention the potential catastrophe that humanity is heading towards, and discuss what needs to be done to avoid it. If we’re to ensure AI remains beneficial to humans in the long run, we may need to radically rethink its design.

Who Should Listen to Human Compatible?

  • AI specialists who could benefit from a new approach to AI design
  • Students of AI looking for an overview of its key problems
  • Anyone concerned about how AI is likely to transform society in the near future

About the Author: Stuart Russell

Stuart Russell is a professor of computer science at the University of California, Berkeley. A leading AI researcher, Russell has served as vice-chair of the World Economic Forum’s Council on AI and Robotics and as an advisor to the UN regarding arms control. He’s also coauthor of the definitive and universally acclaimed textbook on AI – Artificial Intelligence: A Modern Approach (1994) – which is the number one best-selling textbook on AI and is used in over 1,400 universities in 128 countries around the world.

🎧
Listen in the AppOffline playback & background play
Get App