Elements of AI: Introduction to AI ↗
Free course · beginner · registration for full access
Start with Chapter 1. Explain one example of AI and one limitation in your own words. No coding background required.
Build your career plan.
Answer a few questions to open your workspace: a timeline, clear actions and useful resources, with progress saved as you go.
About two minutes · No account needed · You can change your answers
Opening your career workspace…
Prefer to read a guide first?
Read the plain-language career guide →
New to AI? Start with these six terms.
- AI
- Software that performs tasks associated with intelligence, such as recognizing patterns or generating an answer.
- Generative AI
- AI that produces text, images or other output from learned patterns and your input. Its answer is not automatically true.
- Prompt
- The instructions, question or material you give a system. Clear inputs help you define the task; they do not guarantee a correct result.
- Agent
- An AI system connected to tools so it can take steps toward a goal. Permission to send, buy, edit or delete needs more care than permission to draft.
- AGI
- Artificial general intelligence. In this guide’s hypothetical scenario: broadly skilled-human-level cognitive capability. Definitions differ; this is not a claim about today’s systems.
- Superintelligence
- In our hypothetical scenario: AI substantially beyond human cognitive performance across domains. Affordability, reliable deployment and physical capabilities are separate questions.
Plain-language working definitions. Read the basics → · AGI, superintelligence and the research limits →
Optional: try a short AI-checking exercise
Can you catch the invented detail?
This fictional exercise illustrates a plausible AI error. It is a teaching example, not a live model test or a test of your employability.
The complete source
“The community workshop opens at 10 a.m. on Saturday. Booking is required. For cancellations, contact the organizer.”
The draft to check
“The workshop opens at 10 a.m. on Saturday. Book ahead; you can cancel at any time for a full refund.”
Choose an answer, then compare the explanation with the source.
Read the answer and a reusable checking routine
The refund claim is unsupported. Check the source, identify the gap, remove the invented detail, and ask the responsible person. An AI-generated citation is also a claim to verify: open it and find the passage that supports the answer.
Try another public paragraph yourself. A good practice prompt is: “Use only the source below. Draft three answers. For each factual claim, point to the supporting sentence. If the source is silent, say what is unknown.” Then check those references yourself.
Looking further ahead: what if AI becomes much more capable?
No one can promise a secure career after AGI or superintelligence. These optional scenarios help you consider uncertainty; you do not need to solve them before taking a useful step today.
03 / Beyond a single forecast
What if the ground shifts?
Change the assumptions. See where the strategy helps—and where individual preparation is not enough. These scenarios are not sequential predictions or an AGI timeline.
Planning scenario · consistent with varied present-day evidence
Uneven adoption
AI improves some tasks, but adoption, cost and reliability remain uneven.
- What puts pressure on the plan
- Your role may change even when its job title survives. Extra output does not automatically mean higher pay.
- A response to consider
- Measure a useful workflow, keep learning its subject matter, and test whether employers reward the improvement.
- Evidence that prompts a review
- Revisit when your employer changes duties, local vacancies change, or your comparison stops showing a benefit.
- Do not assume
- Assuming every job disappears—or that early aggregate stability protects your job.
Editorial stress tests informed by International AI Safety Report 2026 and the Levels of AGI framework. No probability, arrival date or safe-profession ranking is implied.
Read all four scenarios together / print reference
Planning scenario · consistent with varied present-day evidence
Uneven adoption
AI improves some tasks, but adoption, cost and reliability remain uneven.
- What puts pressure on the plan
- Your role may change even when its job title survives. Extra output does not automatically mean higher pay.
- A response to consider
- Measure a useful workflow, keep learning its subject matter, and test whether employers reward the improvement.
- Evidence that prompts a review
- Revisit when your employer changes duties, local vacancies change, or your comparison stops showing a benefit.
- Do not assume
- Assuming every job disappears—or that early aggregate stability protects your job.
Planning scenario · future reliability is uncertain
More capable agents
Systems become able to perform longer multi-step workflows with less supervision.
- What puts pressure on the plan
- Work based only on drafting, coordinating or checking routine outputs may shrink too.
- A response to consider
- Test an adjacent problem with real demand. Learn to define acceptance checks, permissions and recovery—but verify that people are still paid to do that work.
- Evidence that prompts a review
- Revisit when repeated real-world trials, including failures and full costs, show a workflow can run with substantially less human input.
- Do not assume
- Calling a demo autonomous employment, or assuming human oversight is a permanent moat.
Hypothetical stress test · no arrival date or probability assigned
AGI-level capability
For this exercise, AI matches skilled human performance across a broad range of cognitive work. Definitions differ; capability is not the same as affordable deployment.
- What puts pressure on the plan
- Many current skill premiums could weaken. New tasks, institutions, costs and customer choices would shape paid opportunities.
- A response to consider
- Check actual demand before retraining. Preserve transferable evidence and support relationships, and engage with workplace transition and public-support arrangements.
- Evidence that prompts a review
- Respond to verified deployment and changes in hiring, pay and responsibilities—not a vendor declaring AGI.
- Do not assume
- A list of guaranteed safe professions, or the claim that everyone can become an AI supervisor.
Hypothetical stress test · no validated career forecast
Superintelligence
For this exercise, AI substantially exceeds human cognitive performance across domains; assume it can also be deployed cheaply and reliably. Those are separate assumptions.
- What puts pressure on the plan
- Training alone might not preserve demand for human labor. Purpose, social contribution and access to income are different problems.
- A response to consider
- No personal career recipe can guarantee security here. Collective choices about access, ownership, worker voice, public services and income support become central; remain connected to people and institutions that can act on them.
- Evidence that prompts a review
- Review concrete changes in livelihoods and access to essential services with trusted local institutions. Do not make irreversible decisions based on a speculative timetable.
- Do not assume
- Selling an 'ASI-proof' course, treating unemployment as a personal skill failure, or presenting any policy as guaranteed to solve the scenario.
Browse all learning and support resources
04 / The learning shelf
Open the next useful thing.
You do not need to finish every resource. Start with one lesson, then use the references when a task calls for them. Learning links are recommendations, not evidence that taking a course leads to a job.
Links and access descriptions checked September 5, 2026. No affiliate links. Resources open in a new tab to keep your draft here; external sites have their own accounts, data policies and accessibility. U.S. career tools are labeled; readers elsewhere should check their local employment service and credential rules.
Building AI ↗
Free course · optional next step · certificate costs extra
Use the non-programming exercises to define a problem and the data it would need. Python exercises are optional for this career plan.
Generative AI risk-management profile ↗
Free PDF · reference, not a beginner course
Look up confabulation and data privacy in Section 2. Use these risks to question a proposed AI workflow; this is not compliance certification.
My Next Move ↗
Free occupation explorer · U.S. data
Compare tasks and preparation for two roles. Interest matches are starting points, not evidence of a vacancy or a good personal fit.
Apprenticeship Job Finder ↗
Free search · U.S. opportunities · JavaScript required
Check location, pay, entry requirements and the provider. A registered occupation tag is not the same as a registered employer or a guaranteed place.
American Job Centers ↗
Public career support · U.S. · local services vary
Ask about job-search help, training eligibility, accessibility and local demand before paying for retraining. Outside the U.S., use your public employment service.
International AI Safety Report 2026 ↗
Free report · advanced reading
Read Sections 1.3 and 2.3.1 for capability uncertainty and labor-market effects. Use them to challenge predictions, not to set an AGI countdown.
Browse professional tools and career support
Professional tools and support
Find help for the next step.
Compare public employment services, research tools, application guides, learning and professional advice. Start with an option that fits your location and budget.
Choose help for the step you are taking. Government services, university guides and professional organizations are listed with their purpose and access conditions. Free options come first; paid advice is optional.
Reviewed September 6, 2026 against provider information. Guides support the methods and tools help you act; the suggested pacing, examples and routes are our editorial choices. These resources do not establish that a plan leads to employment. No affiliate links. External links open in a new tab.
Income, benefits and immediate support5 options
Skills, vacancies and training decisions5 options
Applications, interviews and offers9 options
Résumé screening and AI: check the source6 options
Practice with AI2 options
Contract work and a second professional opinion4 options
Options outside the United States2 options
Before booking paid help, agree on scope, relevant experience, credentials, fees and cancellation terms. University guides are public; university advising may require student or alumni eligibility. Outside the listed countries, use your public employment service and local professional bodies.