Why AI predictions are so hard

2026-01-06 · Source: MIT Technology Review · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Fundamental Awareness, short

Summary

MIT Technology Review has published its 2026 AI predictions, acknowledging the increasing difficulty in forecasting the technology's impact despite a strong track record. The article highlights three major unanswered questions complicating AI predictions: whether large language models (LLMs) will continue their incremental intelligence growth, the public's overwhelmingly negative perception of AI, and the confused and fragmented response from lawmakers regarding regulation. It notes public opposition to large data center projects, even those supported by political figures like former President Trump, and the diverse regulatory approaches from entities ranging from California lawmakers to the Federal Trade Commission. While older AI forms like machine learning and deep learning (e.g., AlphaFold) have demonstrated significant "good" applications in science and medicine, the track record for newer LLM-based chatbots is more mixed, showing utility in summarization but also risks like misdiagnosis and unverified claims of discovery.

Key takeaway

For policymakers and technology leaders navigating the AI landscape, understanding the public's negative sentiment and the fragmented regulatory environment is crucial. Your strategic decisions regarding AI development and deployment must account for these external pressures, especially when planning infrastructure like data centers. Proactive engagement with public concerns and collaborative efforts to shape coherent regulation will be essential to mitigate risks and foster sustainable AI progress.

Key insights

Predicting AI's future is challenging due to LLM intelligence plateaus, public unpopularity, and regulatory confusion.

Principles

• Public perception significantly impacts AI adoption.
• Regulatory fragmentation hinders AI governance.

In practice

• Monitor LLM intelligence growth for strategic planning.
• Assess public sentiment for AI project viability.

Topics

• AI Predictions
• Large Language Models
• AI Regulation
• Public Perception of AI
• AI Applications

Best for: General Interest, Policy Maker, Tech Journalist

Related on AIssential

• Poll shows 71% of Americans think AI is progressing too quickly — The rapid pace of AI development is causing widespread public concern over job displacement, transparency, and regulation.
• Why the world needs the UN to keep an eye on AI — Global AI governance is fragmented, with the UN seeking to establish common ground amidst diverse national and private sector approaches.
• 😼 AI's Biggest Problem Just Changed. Nobody's Ready. — AI's primary challenge has shifted from capability to governance and control over autonomous agents.
• I checked out one of the biggest anti-AI protests ever — Growing public concern over AI's potential harms is escalating into organized, large-scale protests.
• AI weather startup outperforms official forecasts — AI's rapid advancement is simultaneously delivering disruptive capabilities and raising urgent safety, ethical, and infrastructure challenges.
• SpaceX Predicts a $26.5T AI Landscape — AI's rapid evolution presents complex challenges in public perception, regulation, market dynamics, and creative integrity.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT Technology Review.