AI as Social Technology

Key Takeaways

• Speculative AGI narratives often mask realistic social impacts of LLMs
• LLMs evolved from translation tools to perceived pathways toward AGI
• Myths about AI's utopian or dystopian futures overlook existing complexities
• Real-world AI effects arise from interaction of imperfect tech and societies
• Mapping AI's social influence now is more actionable than predicting singular outcomes

Pulse Analysis

The rise of large language models has sparked a wave of grandiose predictions that treat AI as a magical catalyst for societal overhaul. Scholars Henry Farrell and Cosma Rohilla Shalizi push back, warning that these narratives borrow from two‑decade‑old myths and obscure the nuanced ways LLMs embed themselves in everyday workflows, from drafting legal contracts to generating code. By stripping away the hyperbole, the authors reveal a technology that is powerful yet fundamentally statistical, capable of mimicking human discourse without possessing true intent or agency.

Technically, LLMs progressed from a modest improvement in machine translation—Vaswani et al.’s 2017 transformer model—to the centerpiece of the AGI debate within a few short years. Their ability to produce coherent text and even functional code has led investors and technologists to label them the "royal road" to general intelligence. However, this leap in perception outpaces empirical evidence; LLMs remain brittle, prone to hallucinations, and heavily dependent on the data they ingest. Recognizing these limits is crucial for businesses that must balance innovation with reliability, especially as AI‑generated content becomes integral to marketing, customer service, and product design.

The authors advocate a pragmatic approach: instead of chasing singular, apocalyptic futures, stakeholders should map AI’s ongoing social impact. This means interdisciplinary research that tracks how LLMs alter labor markets, decision‑making hierarchies, and regulatory compliance. Policymakers can then craft targeted guidelines—such as transparency standards for AI‑generated outputs—while firms develop internal audit frameworks to monitor bias and performance drift. By treating AI as a social technology rather than a mythic force, the industry can navigate uncertainty with evidence‑based strategies, turning potential disruption into a manageable, competitive advantage.