Why does Sam Harris believe superhuman AI poses inherent danger?

Sam Harris argues that there's inherent danger for the 'dumber party' in a relationship with a smarter entity, drawing an analogy to humans and dogs. He believes our inability to fully comprehend a superior intelligence makes alignment with human well-being incredibly difficult, as AI might pursue goals we cannot conceive or might not align with our values.

What is the 'superhuman calculator' analogy concerning narrow AI?

The 'superhuman calculator' analogy highlights that even narrow AI today, like a phone's calculator, is superior to humans in its specific task. This implies that once AI achieves general intelligence, it won't be human-level for long, but immediately superhuman, rapidly improving beyond our comprehension.

What are Sam Harris's immediate concerns about the impact of narrow AI?

Harris is deeply worried that within 12-24 months, generative AI like GPT-5 will make most online information fake, creating a 'cacophony of misinformation.' This could undermine trust in institutions, lead to political instability, and make it impossible to discern truth, potentially requiring a 'bankruptcy of the internet'.

How did Twitter impact Sam Harris personally?

Sam Harris describes Twitter as a 'chaos monster' that amplified misleading signals and caused significant interpersonal and professional issues for him over a decade. He found his engagement with the platform to be a 'psychological experiment' that distorted his view of humanity, leading to immense relief after deleting his account.

What solution does Sam Harris propose for AI-driven misinformation?

Harris suggests a new kind of 'gatekeeping' based on digital provenance. He envisions a future where information, especially videos and images, will require verification from trusted sources (like news organizations or technical watermarks) before being accepted as real, shifting the burden of proof to authenticity.

What was Elon Musk's response when asked about purpose in an AGI future?

When asked how he'd advise his children on purpose in a world where AGI is near, Elon Musk hesitated for 15 seconds, stuttered, and eventually stated he lives in 'suspended disbelief.' He questioned the point of his own work if AGI fundamentally changes all human endeavors.

How does Sam Harris suggest we find meaning and purpose in an AI-dominated future?

Harris proposes a return to the humanities, focusing on art, philosophy, and understanding what it means to live a good life. He argues that while AI can automate many technical jobs, the value of conscious human creation and interpersonal dialogue in fields like art, literature, and philosophy will remain paramount.

What is Sam Harris's take on Universal Basic Income (UBI) in an AI-enabled economy?

Harris believes that if AI successfully produces abundant goods and cancels the need for much human labor, UBI would become necessary. He argues society would need to redefine the value of work, decouple survival from labor, and find new ways for people to pursue purpose and happiness in a post-scarcity world.

What is the transformative power of radical honesty, according to Sam Harris?

According to Harris, choosing not to lie (even white lies) profoundly changes one's life by eliminating a 'gateway to misbehavior.' It fosters authentic relationships, provides valuable feedback, and requires individuals to confront uncomfortable truths about themselves and others, ultimately boosting happiness and preventing reputational damage.

What role do psychedelics and meditation play in Sam Harris's philosophy?

Harris sees psychedelics and meditation as fundamental tools for self-insight, particularly useful for rites of passage and approaching death wisely. He advocates for structured environments, akin to ancient mysteries, that facilitate these experiences to cultivate wisdom and spiritual understanding in a secular context.

What is the Waking Up app and why is it important to Sam Harris?

The Waking Up app is a platform for guided meditation and teachings on secular spirituality, expanding on themes from his book. Harris considers it a perfect delivery system for such information, which helps users focus attention and reorder priorities, countering the fragmenting effects of typical smartphone usage.

Key Moments

WARNING: ChatGPT Could Be The Start Of The End! Sam Harris

The Diary Of A CEO

People & Blogs4 min read111 min video

Aug 7, 2023|2,161,699 views|41,883|7,589

The Diary Of A CEO steven bartlett steve bartlett podcast the diary of a CEO podcast life lessons CEO sam Harris interview Who will AI replace?AI Could Destroy Us AI could destroy the internet how to be happy can we build AI without losing control over it?

Save to Pod

Key Moments

TL;DR

Sam Harris discusses AI risks: misinformation, alignment with AGI, and societal upheaval.

Key Insights

AI poses both near-term risks (misinformation, societal disruption) and long-term existential risks (AGI misalignment).

The rapid, unconstrained development of AI, especially AGI, presents a significant danger due to the intelligence mismatch between humans and machines.

The internet's usability is threatened by AI-generated misinformation, potentially destabilizing societies and undermining democratic processes.

Humanity's incentives to develop AI, driven by its immense value, make it difficult to pause or control its progress.

A shift towards the humanities and ethical considerations is crucial to navigate a future increasingly shaped by advanced AI and to find human purpose.

Universal Basic Income (UBI) is presented as a potential solution to widespread job displacement caused by AI, necessitating a reevaluation of work and value.

Honesty and a willingness to change one's mind, particularly in the face of truth and diverse perspectives, are essential for individual well-being and societal progress.

THE DUAL THREAT OF ARTIFICIAL INTELLIGENCE

Sam Harris outlines two primary concerns regarding artificial intelligence: the near-term issues stemming from human misuse of increasingly powerful AI, and the long-term existential threat posed by Artificial General Intelligence (AGI) that surpasses human capabilities. The near-term dangers include the amplification of misinformation and disinformation, making it increasingly difficult to discern reality and potentially destabilizing democratic processes like elections. The long-term concern revolves around the 'alignment problem,' ensuring that superintelligent AGI remains aligned with human interests once developed, a challenge many fail to adequately grasp.

THE INEVITABILITY AND DANGER OF SUPERHUMAN AI

Harris argues that the development of superhuman AI is nearly inevitable, based on two core assumptions: intelligence is substrate-independent (not requiring biological matter) and that progress in AI development will continue due to immense incentives. He likens the arrival of superintelligent AI to an alien species landing on Earth, emphasizing the inherent danger of a less competent species being in the presence of a vastly superior one. This intelligence mismatch means humans may not understand the motives or actions of advanced AI, leading to potentially catastrophic outcomes without malice, akin to how humans treat less intelligent species.

THE INTERNET'S EROSION AND SOCIETAL FRAGMENTATION

A significant near-term risk highlighted is the potential for AI to render the internet unusable due to mass generation of fake information. Advanced AI could soon produce convincing fake texts, images, and videos at an unprecedented scale, making it nearly impossible to distinguish truth from falsehood. This erosion of trust in online information, combined with existing societal divisions amplified by social media, could lead to political instability, breakdown of cooperation, and an inability to conduct valid democratic elections. Harris draws a parallel to personal negative experiences on platforms like Twitter, which he views as prioritizing engagement over truth and contributing to psychological distress.

THE CHALLENGE OF ALIGNMENT AND UNFORESEEN CONSEQUENCES

The core of the long-term AI risk is the alignment problem – ensuring AI goals remain aligned with human values as it evolves. Harris challenges the optimistic view that greater intelligence inherently leads to greater ethics, suggesting a vast space of possible superhuman intelligences, many of which may not be benign. The rapid, uncontained deployment of AI, like its integration into existing systems before safety considerations are fully addressed, bypasses crucial checkpoints. The incentives to advance AI development, even with awareness of risks, make a global pause or control extremely difficult, setting the stage for potentially irreversible outcomes.

REDISCOVERING HUMANITY AND THE VALUE OF PURPOSE

In an age of advancing AI, Harris suggests a return to the humanities – philosophy, art, and the exploration of what it means to live a good life – as these are the areas least likely to be automated and most central to human experience. He contrasts this with professions directly threatened by AI, like software engineering. The conversation touches on finding purpose and meaning, drawing on Elon Musk's admission of "suspended disbelief" regarding AI's ultimate point. Harris emphasizes the intrinsic value of human connection and consciousness, arguing that even with AI-generated content, the human source in creative and philosophical endeavors holds unique importance.

RETHINKING ECONOMICS AND THE FUTURE OF WORK

The potential for AI to displace human labor across many sectors, including high-status cognitive jobs, necessitates reconsideration of economic models. Harris posits that AI might replace jobs without creating equivalent new ones, unlike previous technological revolutions. This could lead to widespread unemployment and social instability unless societal structures adapt. Universal Basic Income (UBI) is discussed as a potential solution to distribute the wealth generated by AI, ensuring survival and allowing individuals to pursue non-work-related activities, hobbies, and relationships, thereby redefining purpose beyond economic necessity.

THE PATH TO CHANGING MINDS AND FOSTERING TRUST

Harris highlights the importance of honesty in personal and societal interactions, arguing that a commitment to truthfulness recalibrates relationships and fosters genuine trust. He criticizes the corrosive effect of white lies and the tendency to dismiss information based on the source rather than its merit. Rebuilding trust in institutions, especially in political and scientific spheres, is crucial for societal coherence. Drawing parallels to the COVID-19 pandemic, he notes society's failure to effectively manage misinformation and division, underscoring the need for improved institutional trustworthiness and a willingness to engage in open, honest dialogue to navigate complex challenges.

Mentioned in This Episode

●Software & Apps

●Companies

●Organizations

●Books

●People Referenced

Common Questions

Six years after his TED Talk, Sam Harris remains pessimistic about AI safety. He's concerned about both near-term issues like misinformation from narrow AI and the long-term alignment problem with artificial general intelligence (AGI). He emphasizes that current AI development is already 'in the wild' without adequate safety protocols.

Mentioned in this video

People

Alex Gibney

An American documentary filmmaker, referred to as an example of a filmmaker whose style could be replicated by advanced AI for misinformation.

Ron Howard

A professor at Stanford who taught a seminar on the ethics of lying, significantly influencing Sam Harris's views.

Companies

Getty Images

An American visual media company, mentioned as an example of a trusted source for verified images, contrasting with generic unverified content online.

Software & Apps

Chrome browser

Google's web browser, used as an example of how technological solutions like digital watermarks could be democratized for users.

Waking Up App

Sam Harris's meditation and mindfulness application, described as an effective delivery system for secular wisdom and a counter to smartphone's fragmenting effects.