AI Fear
Fear AI
An early illustrated post. A perspective on our relationship with AI, rather than a measurement of its effects.
Home / Sources
THE SOURCE ROOM
Sources and
references
A short reading list with the qualifications left in.
Research, reporting and opinion do different jobs.
7 sources
International Energy Agency
Key Questions on Energy and AI
Separate historical estimates from scenarios, and all data-center electricity from AI-focused facilities. See Figure 1.5, pp.25–26.
MIT researchers
Beyond AI Exposure
Svanberg and coauthors model the cost of automating computer-vision tasks. Under their assumptions, economically automatable tasks account for 23% of compensation for vision tasks (abstract, p. 1). The MIT–IBM Watson AI Lab funded the study.
The Verge / Vox Media
Hope, fear, and AI
An early public-attitudes survey and reported feature. Read it as a historical snapshot, not today’s view of every AI system.
Rex Woodbury / Digital Native
How to Productize AI Anxiety
Investor commentary on anxiety, incentives and product positioning. The commercial perspective is part of the context.
Ed Lyons
Software, in a Time of Fear
A first-person perspective on software practice. Personal experience can illuminate a question without establishing an economy-wide result.
World Economic Forum
The Future of Jobs Report
Employer expectations and scenarios. Keep expected task and skill changes distinct from observed employment outcomes.
HOW WE READ
Evaluating
sources
Attach a source to the particular fact it supports. Preserve the document name, date, population, unit and page or table. Say when a result is an estimate, a forecast, an opinion or an observation.
Authority is claim-specific. A company can know its operations and have an interest in presenting them favorably. A study can be methodical and still have limited coverage. More headlines do not necessarily mean more independent evidence.