Source-linked AI summary
The Turing Trap: The Promise & Peril of Human-Like Artificial Intelligence
Erik Brynjolfsson
TL;DR
The paper examines the distinction between AI as a tool and AI as a mind, and the consequences of directing technology toward augmenting or replacing labor. It argues that broad training and tax incentives can let markets reward labor-augmenting approaches, while acknowledging that substitution and complementarity are difficult to distinguish in practice.
Problem
The paper addresses how artificial intelligence should be understood and directed, distinguishing computers as powerful tools from systems that “really” are minds.
Method
The paper uses a conceptual distinction between labor substitution and augmentation to motivate broad training and tax incentives rather than technology-specific mandates or prohibitions.
Results
Markets can ultimately reward approaches that augment labor rather than replace it.
Takeaways & Limitations
Broad training and tax incentives allow technologies, entrepreneurs, and markets to reward labor-augmenting approaches.
Takeaways & Limitations
The distinction between labor complements and substitutes is clear in theory but trickier to apply in practice.
Abstract
from arXiv · showhide
In 1950, Alan Turing proposed an imitation game as the ultimate test of whether a machine was intelligent: could a machine imitate a human so well that its answers to questions indistinguishable from a human. Ever since, creating intelligence that matches human intelligence has implicitly or explicitly been the goal of thousands of researchers, engineers, and entrepreneurs. The benefits of human-like artificial intelligence (HLAI) include soaring productivity, increased leisure, and perhaps most profoundly, a better understanding of our own minds. But not all types of AI are human-like. In fact, many of the most powerful systems are very different from humans. So an excessive focus on developing and deploying HLAI can lead us into a trap. As machines become better substitutes for human labor, workers lose economic and political bargaining power and become increasingly dependent on those who control the technology. In contrast, when AI is focused on augmenting humans rather than mimicking them, then humans retain the power to insist on a share of the value created. Furthermore, augmentation creates new capabilities and new products and services, ultimately generating far more value than merely human-like AI. While both types of AI can be enormously beneficial, there are currently excess incentives for automation rather than augmentation among technologists, business executives, and policymakers.
7 Erik Brynjolfsson and Andrew McAfee, “The Business of Artificial Intelligence,” Harvard
The section situates human-like AI within debates about machine intelligence and distinguishes strong from weak AI. It also notes a theoretically possible but empirically unlikely risk that living standards could fall as productivity rises.
- Human-like AI: Strong AI is described as computer intelligence that “really is a mind,” whereas weak AI treats computers as powerful tools.The distinction is attributed to John Searle.
- Human-like AI: Creating human-matching intelligence has been framed by some researchers as computer science’s “manifest destiny.”This claim appears in the discussion of strong and weak AI.
- Risks: Living standards could theoretically fall even as productivity rises if working hours decline fast enough.The passage characterizes this outcome as empirically unlikely.
15 See for example Daron Acemoglu, “Directed Technical Change,” Review of Economic
The section discusses how technological change can substitute for or complement labor, while emphasizing that the distinction is difficult to apply in practice. It favors broad training and tax incentives that allow markets to reward technologies augmenting rather than replacing labor.
- Technological change: Digital photography expanded from an estimated 85 billion photos in 2000 to 1.4 trillion in 2020.The later figure was almost entirely digital and represented nearly twenty-fold growth.
- Labor and technology: Technological change can substitute for labor or complement it, but distinguishing these effects is trickier in practice than in economic theory.The passage presents this difficulty as a limitation on applying the theoretical distinction.
- Policy: Broad training and tax incentives are appealing because they avoid specific technology mandates or prohibitions.These instruments leave room for technologies, entrepreneurs, and markets to determine which approaches succeed.
- Policy: Markets can ultimately reward approaches that augment labor rather than replace it.The passage presents this as the intended direction enabled by broad policy tools.