AI Economy Quiz: How Ready Are You for the Transformation Ahead?

Why does research suggest 'learn by doing' outperforms passive learning for AI skill development?

Hands-on experimentation builds task-specific heuristics (when to trust AI, where it fails) that passive content can't provide — practice generates calibration that reading about AI doesn't Passive learning about AI is ineffective because AI hasn't been studied enough yet Workers learn better by doing for all subjects equally — there's nothing AI-specific about this AI courses and certifications are lower quality than self-directed experimentation

Why is maintaining deep domain expertise emphasized alongside developing AI fluency, rather than prioritizing one over the other?

AI fluency without domain expertise produces volume without quality; domain expertise without AI fluency loses productivity to AI-augmented competitors. The combination — T-shaped AI-fluent experts — produces the highest-quality output at scale Domain expertise is no longer valuable once AI can perform most knowledge work tasks Only AI engineers need domain expertise; most professionals only need basic AI tool fluency AI fluency is sufficient — domain expertise is what AI provides from its training data

A company plans to run a customer service AI chatbot handling 500,000 short conversations per month. Which cost dimension should they optimize first?

Always use the highest-capability model to ensure quality Choose the model with the largest context window Pick the model with the best score on the MMLU benchmark Use a smaller, faster, cheaper model (e.g., Haiku or Flash tier) since the tasks are short and high-volume

A company uses AI to screen resumes and notice that the tool systematically ranks candidates from certain universities higher. What is the most likely cause?

There is no cause — AI resume screening is required to be unbiased by design The AI was trained on historical hiring data where those universities were over-represented among successful hires — encoding past selection bias The AI is correctly identifying the objectively better universities The university ranking reflects the AI's access to university prestige databases

Why is 'one and done' thinking about AI systems a failure pattern?

AI systems are one-time costs and need no maintenance after deployment Organizations should replace AI systems with new ones every year as a best practice One and done thinking only fails for AI systems used in consumer products, not internal tools AI systems require ongoing monitoring and updating — the underlying model providers update their systems, data distributions shift, and usage patterns change in ways that can degrade AI performance over time

A restaurant receives 20 new Google Reviews per week. Responding to all of them takes too much time. What is the best AI-assisted approach?

Use AI to respond automatically to all reviews without any human review Consolidate all responses into one weekly post Respond only to negative reviews — positive reviews don't need responses Use AI to generate personalized response drafts for each review that the owner quickly reviews and publishes

Which statement about AI API output tokens is correct?

Output tokens are not counted toward API costs at the Sonnet/GPT-4o tier Output tokens are always cheaper because the model is just retrieving stored answers Input and output tokens are always priced identically at major providers Output tokens are typically more expensive per token than input tokens

Why does critical thinking become MORE valuable when AI handles more information work?

AI eliminates the need for critical thinking by providing verified correct answers Critical thinking only matters for AI engineers and researchers, not everyday professionals Critical thinking is primarily for detecting AI hallucinations and has little other value AI generates more options, analyses, and drafts than humans could previously produce — but doesn't reliably know which option is best. Choosing wisely among AI-generated options becomes the bottleneck, amplifying the value of good judgment

What is the purpose of a 'personal AI exposure audit' in career planning?

It is primarily used to calculate how much of your salary should be replaced by AI tools An audit shows which AI tools are approved for use in your organization A personal AI audit is required by employers to certify workers understand AI policies To categorize your current tasks by AI substitutability and AI amplification potential — so you can focus energy on skills and tasks where investment has the highest career return

What is meant by 'AI governance' in an organizational context?

AI governance means ensuring the AI system is connected to the internet for current information AI governance refers only to government regulation of AI companies AI governance is the technical process of fine-tuning AI models The structures and processes that establish accountability for AI use: who approves high-risk uses, who is responsible when AI produces errors, how AI output quality is monitored, and what uses are prohibited

For a worker in a high AI substitution risk role, what is the most effective adaptation strategy?

Avoid using AI tools entirely so that your role remains necessary Specialize more deeply in the exact tasks AI is replacing to become the best human doing those tasks Advocate for regulation that bans AI in your industry Upskill into adjacent roles with more human judgment content — a data entry clerk learning SQL and data analysis moves into an AI-augmented growth role rather than a declining substituted one

What is meant by 'distributional tension' in the context of AI economic gains?

AI companies are required to distribute profits to all affected workers AI creates equal productivity gains for all workers regardless of skill level AI productivity gains are initially concentrated among AI company shareholders and skilled tech workers, raising concerns about inequality before benefits spread broadly Distributional tension refers to AI models that provide uneven answer quality

Which AI capability allows a small business to maximize the reach of a single piece of content?

Using AI to repurpose one piece of content into multiple formats — blog to social posts, email, FAQ, LinkedIn article Having AI scrape competitor content and rewrite it Using AI to automatically post the same content to all platforms simultaneously AI translating content into multiple languages automatically

Why are skilled trades like electricians and plumbers considered AI-resilient occupations?

They require physical dexterity in variable, unpredictable physical environments — something current AI cannot perform — and industry aging plus declining new entrants is creating wage-driving shortages AI models already work as plumbers and electricians in smart homes Skilled trades require no judgment and are therefore not worth automating Skilled trades are fully automated but workers are still needed to supervise the machines

A bakery owner wants to post daily on Instagram and send a weekly email newsletter but has no time to write content. What is the most practical AI solution?

Use AI (ChatGPT or Jasper) to generate captions and email drafts from brief descriptions of weekly specials, saving hours of writing each week Wait for platform-native AI tools directly in Instagram Use AI to autonomously post to Instagram without any human review Hire a social media agency — AI is not ready for food business content

A consultant generates proposals manually, which takes 3 hours each. They close 30% of proposals. What is the most direct AI impact opportunity?

Use AI to find and pitch leads, bypassing the proposal stage Use AI to generate first-draft proposals from a project brief, reducing the 3-hour task to 1 hour of review and customization — enabling more proposals at the same time investment Use AI to increase the proposal price to improve ROI per proposal Use AI to automatically send proposals without owner review to maximize volume

Which category of work is MOST protected from AI task displacement according to current research?

Information processing and synthesis, which is well-represented in AI training data Language production tasks like email composition and first-draft writing Pattern recognition in structured, well-defined datasets Physical dexterity in variable, unpredictable environments — like plumbing or electrical work — because AI models can't take physical action and unstructured physical contexts are difficult to automate

What types of AI use are 'banned outright' under the EU AI Act's 'unacceptable risk' category?

Social scoring by governments and manipulation of vulnerable groups — AI that systematically rates citizens' social behavior or exploits psychological vulnerabilities are prohibited as fundamentally incompatible with rights Any AI system that has ever produced an incorrect output is classified as unacceptable risk All AI used in healthcare and finance is prohibited under the EU AI Act AI systems costing more than €1 million to develop are automatically classified as unacceptable risk

What is the specific goal for Days 31–60 in the 90-day personal action plan?

Build an AI-assisted workflow for one recurring time-consuming task that reduces the time for that task by at least 30%, and experiment with one adjacent AI tool category Days 31–60 are for rest — consolidating what was learned in the first 30 days Replace your entire current workflow with AI-generated processes Complete an AI certification course to demonstrate formal AI proficiency

What is the EU AI Act's 'high risk' category, and what does it require?

The EU AI Act's high-risk category only applies to AI systems built in Europe, not imported High-risk AI systems require approval from the AI system's provider but no additional organizational obligations High-risk AI systems (used in hiring, credit, healthcare, and critical infrastructure) must meet strict requirements for transparency, human oversight, accuracy, and safety before deployment High-risk AI is banned outright in the EU — organizations cannot use AI in these domains