AI Skills for Jobs: What Young Workers Need and How to Start

What are AI skills for jobs, and which ones do employers want now?

AI skills are the abilities to use, judge and build artificial intelligence systems for real work. AI literacy is the basic version: knowing what AI tools can and cannot do, using them sensibly and checking what they produce. Employers talk about both, and the main sources measure them in different ways.

What do employers say?

The World Economic Forum's Future of Jobs Report 2025, published on 7 January 2025, draws on more than 1,000 employers representing over 14 million workers across 22 industry clusters and 55 economies, and covers the period 2025-2030 ^[1]. It follows the 2023 and 2020 editions ^[1]. In it, AI and big data top the list of fastest-growing skills, followed closely by networks and cybersecurity and technology literacy ^[1].

Hiring intent points the same way. Two-thirds of surveyed employers plan to hire talent with specific AI skills, while 40% anticipate reducing their workforce where AI can automate tasks ^[1]. For a young worker both numbers matter: demand for AI skills is rising, and so is pressure on the tasks AI can already do.

What does LinkedIn's data show?

LinkedIn's 2026 Skills on the Rise list for the US, as reported by Quartz, has 49 skills in eight areas ^[2]. The largest group is a set of 10 AI-building skills. Others include AI strategy skills such as AI for Business, Responsible AI and Data Governance, and people skills such as public speaking and stakeholder management ^[2]. Prompt engineering and data annotation appear as technical examples ^[2].

Keep the source types apart. The WEF figures come from an employer survey across many economies. The LinkedIn list is platform data for one country. Neither is a promise that a particular job exists for you.

What AI skills should I learn first?

None of the sources reviewed offers one official curriculum. The three tiers below are an editorial way of grouping what they describe, from the easiest starting point to the most technical.

Tier 1: everyday use

Fortune reported on 21 March 2025 that LinkedIn ranked AI literacy as the fastest-growing skill for 2025, with soft skills right behind ^[3]. In practice this means using generative AI tools, which produce text, images or code from a prompt, for drafting, summarising and research, and getting into the habit of checking what comes back against a reliable source.

Tier 2: work skills around AI

Analytical thinking remains the most sought-after core skill, considered essential by seven out of 10 companies in 2025 ^[1]. LinkedIn's list also includes Responsible AI and Data Governance ^[2]. Reading a spreadsheet, questioning a number and spotting a weak source are the everyday forms of these skills.

Tier 3: technical skills

Machine learning, prompt engineering, data annotation and other AI-building skills sit here ^[2]. The WEF finds that technology-related roles, including big data specialists, fintech engineers, AI and machine learning specialists and software developers, are the fastest-growing jobs in percentage terms ^[1]. Not everyone needs this tier, but it is where the fastest percentage growth is.

Do I need to learn to code?

The sources do not say coding is a requirement. Sam Altman compared getting good at AI tools to getting really good at coding when he finished high school ^[3], and software roles are among the fastest-growing ^[1]. Yet much of LinkedIn's list, such as AI for Business, Responsible AI and stakeholder management, is not programming ^[2]. Choose based on the roles you want.

What is Sam Altman's advice for young workers?

Two pieces of advice are on the record in the sources, 19 months apart.

March 2025: learn to use the tools

Altman told Stratechery's Ben Thompson, as Fortune reported on 21 March 2025, that the obvious tactical thing is to 'get really good at using AI tools', and he compared it to getting really good at coding when he was graduating from high school ^[3]. Fortune also mentions a months-long skills training programme as one stepping stone ^[3].

October 2026: try something new

In a Stanford Graduate School of Business 'View From The Top' interview released on 21 October 2026, Altman said huge numbers of tasks will be done better, cheaper and faster by AI. He cautioned that the change will not come as fast as many expect, and told students 'this is a great time to go do the new thing', according to TheStreet on 29 October 2026 ^[4]. The same article cites US Census Bureau data showing 5.7 million new business applications in 2025, a decade high, and a survey in which 51% of Gen Z had seriously considered starting a business in the past 12 months versus 27% of all Americans surveyed ^[4]. These are secondary reports of the data.

How should you weigh it?

Altman leads OpenAI, which sells AI products, so his advice is one voice among several. It also fits the other evidence. The WEF puts analytical thinking first, alongside resilience and lifelong learning ^[1], and LinkedIn pairs technical AI skills with people skills ^[2]. A reasonable reading is: use the tools, but do not stop at the tools.

Is AI taking entry-level jobs?

There are signs that hiring of young workers is weaker in some AI-exposed occupations, but the evidence is early, mostly American and disputed. An entry-level job here means a first role for a worker aged roughly 22-25, the group the main studies track.

What does the Stanford research find?

Using ADP payroll data on millions of US workers through June 2026, Stanford researchers Brynjolfsson, Chandar and Chen find that employment of workers aged 22-25 in AI-exposed occupations is 19% below where it would be had it kept pace with less-exposed peers. Experienced workers show no comparable gap ^[5]. The paper finds no evidence of widespread, economy-wide job displacement. It says the gap has widened steadily since it was first documented in August 2025, and that it works mainly through reduced hiring of young workers rather than more separations ^[5].

The declines are concentrated in occupations where AI usage mainly substitutes for human tasks. Where AI mainly complements workers, employment is flat or rising, especially for experienced workers ^[5].

What other evidence points the same way?

The Stanford AI Index 2026, as summarised by United Nations University, reports that employment of software developers aged 22 to 25 has fallen nearly 20% since 2024, and that a third of surveyed employers expect workforce reductions in the year ahead, in service operations, supply chain and software engineering ^[6]. TheStreet, citing a Harvard working paper, reports that firms adopting generative AI cut junior employment by about 7.7% within six quarters while senior jobs held steady ^[4]. That is a secondary report and the primary paper was not checked for this article, so treat the figure as unverified.

What do sceptics say?

The OECD Employment Outlook 2026 (July 2026) looks at unemployment among young labour market entrants with and without a college degree. It says unemployment has risen for young people, but that recent AI advances appear to explain little of it so far; it points instead to young entrants' exposure to the labour market slowdown and to longer-term shifts in technology and skills ^[7].

The Stanford authors are cautious too. They call their results early, descriptive indicators rather than causal estimates. The patterns weaken when controlling for education, show some divergent trends before generative AI, and are stronger in the ADP sample than in national survey benchmarks ^[5].

Which jobs are most exposed to AI, and what does that mean?

Occupational AI exposure is a measure of how many of a job's tasks AI could do or help with. It is not a forecast that the job vanishes.

How many workers are exposed?

The International Labour Organization's May 2025 update estimates that one in four workers worldwide are in an occupation with some degree of generative AI exposure. Because human input is still needed, it concludes that most jobs will be transformed rather than made redundant ^[8]. The ILO scores nearly 30,000 tasks, and its mean automation score was 0.29 in 2025 against 0.30 in 2023 ^[8].

Why is exposure not the same as risk?

An OECD paper, Skills in the AI age (July 2026), states that progress in AI capabilities is not the same as the probability of automation. Its figure titles say high-skill white-collar occupations are among those with the highest AI exposure, and that the occupations most at risk of automation differ from those most exposed to AI ^[9]. A well-paid analyst job can be highly exposed because AI can help with many of its tasks, without being the job most likely to be automated away.

Which roles grow and shrink?

The WEF finds technology-related roles are the fastest-growing in percentage terms, while clerical and secretarial roles are expected to see the largest absolute decline ^[1]. Its survey implies that by 2030 about 170 million jobs are created (14% of today's employment) and 92 million displaced (8%), a net gain of 78 million (7%) ^[1]. Employers expect 39% of workers' existing skill sets to be transformed or become outdated over 2025-2030, down from 44% in the 2023 edition and 57% in 2020 ^[1]. Skills change is expected to continue, but employers see it as slower than in earlier surveys.

Do AI skills really pay more?

One large dataset says yes, with a warning attached. A wage premium is the extra pay workers with a skill earn compared with similar workers without it.

PwC's 2025 Global AI Jobs Barometer reports that workers with AI skills earned an average 56% wage premium in 2024, up from 25% the year before ^[10]. PwC's examples of AI skills are machine learning and prompt engineering ^[11]. It also says jobs requiring AI skills grew 7.5% year on year while total job postings fell 11.3% ^[10].

How far should you trust the figure?

PwC's own footnote cautions that the premium may partly reflect the scarcity of workers with AI skills, not only the value of the skills themselves ^[11]. Scarcity tends to fade as more people learn the skills. Both numbers are also averages across many workers and postings, so they say little about what one graduate would be offered.

The picture also sits alongside the entry-level research. PwC finds jobs growing even where tasks are easy to automate ^[10], while Stanford finds falling hiring of young workers where AI substitutes for tasks ^[5]. Both can be true at once: demand for people who can use AI can rise while the number of starter roles in some exposed occupations shrinks.

Which human skills still matter alongside AI?

Employer surveys keep putting human skills next to technical ones. The WEF says analytical thinking is the most sought-after core skill, essential for seven out of 10 companies in 2025, and expects creative thinking, resilience, flexibility and agility, and curiosity and lifelong learning to rise in importance ^[1].

The same report shows why learning never stops. 63% of employers name skill gaps as a major barrier to business transformation over 2025-2030, and 85% plan to prioritise upskilling their workforce ^[1]. Upskilling means learning new skills for your current or a closely related job.

LinkedIn's 2026 US list pairs its AI-building skills with people skills such as public speaking and stakeholder management ^[2]. Fortune reported that LinkedIn's 2025 ranking put soft skills right behind AI literacy ^[3].

Could leaning on AI weaken your own skills?

The Stanford AI Index 2026, as summarised by United Nations University, found that 73% of experts expect AI to improve how people do their jobs versus 23% of the public. It also flagged a concern that heavy reliance on AI tools may carry a long-term learning cost ^[6]. The practical lesson for early-career workers is to use AI to go faster at tasks you understand, and to keep doing some of the thinking yourself.

How can a beginner start building AI skills?

The steps below mix sourced facts with general editorial advice. This is general information, not personal career or financial advice. For decisions that involve paying for courses or changing direction, talk to a qualified careers adviser.

1. Use AI tools on real tasks. Practise on work you already understand. Fortune also mentions a months-long skills training programme as one stepping stone ^[3].
2. Check every output. Compare answers with a reliable source. This protects against errors and keeps your own skills sharp, which matters given the learning-cost concern ^[6].
3. Build analytical and data basics. Analytical thinking is the most sought-after core skill ^[1].
4. Look at how AI is used in your target field. Declines for young workers concentrate where AI mainly substitutes for tasks, while employment is flat or rising where it complements workers ^[5].
5. Show your people skills. Public speaking and stakeholder management are on LinkedIn's 2026 list ^[2].

What might a portfolio look like?

This example is illustrative, not drawn from a source. A graduate applying for an operations role keeps three short case notes. In the first, she summarises a long report with an AI tool and lists the three facts she verified. In the second, she cleans a messy spreadsheet and explains one number she questioned. In the third, she presents the results to a mock client. Together they show tool use, checking and communication, which are the three things the sources keep pairing.

What do AI-first employers say?

Microsoft's 2025 Work Trend Index surveyed 31,000 workers across 31 countries. It found 78% of leaders are considering hiring for AI-specific roles, rising to 95% at 'Frontier Firms', Microsoft's term for AI-first companies ^[12][13]. Microsoft also says that in Frontier Firms even entry-level employees are managers from day one because they manage AI ^[12]. That is a claim from a company that sells AI products, not independent evidence.

What is different for readers in Iran?

Access is the main difference, and it changes often. IranWire reported in October 2026 that access to ChatGPT, Claude and Grok has been disrupted for users in Iran connecting without a VPN. It adds that Iranian users have long lacked direct access because of restrictions by foreign companies and US sanctions ^[14].

The same report says infrastructure-level routing had recently allowed VPN-free access, and that ChatGPT was unavailable without a VPN in August 2026 before being restored after a few days ^[14]. It does not make clear whether Iranian infrastructure or foreign providers caused the latest disruption ^[14]. IranWire also describes AI tools as part of daily work for students, researchers, programmers, journalists and online business operators in Iran ^[14].

What can you build that does not depend on one tool?

The most durable skills are the ones that move between tools. Analytical thinking ranks first among core skills in the WEF survey ^[1], and LinkedIn's 2026 list includes people skills such as public speaking and stakeholder management ^[2]. If one tool is unavailable, skills like these still carry over to the next one.

This research round found no credible source on Iranian hiring data, payment limits for AI subscriptions or local AI alternatives, so this article does not guess at them. It also does not give instructions for evading restrictions. Check each provider's terms of service and the rules that apply to you before relying on any tool for paid work.

What is still uncertain about AI and young workers' careers?

Four open questions matter most for anyone planning an early career.

Is the entry-level effect caused by AI?

Not yet established. Stanford calls its findings descriptive rather than causal and notes that patterns weaken when controlling for education ^[5]. The OECD points instead to the labour market slowdown and longer-term shifts in technology and skills ^[7]. Both can be partly right.

How fast will change arrive?

Altman says the transformation will be slower than many expect ^[4]. Yet Stanford's gap has widened steadily since August 2025 ^[5], and a third of employers surveyed for the AI Index expect workforce reductions in the year ahead ^[6]. Employers in the WEF survey also expect less skills disruption (39%) than in 2023 (44%) and 2020 (57%) ^[1].

Does the evidence apply outside the US?

The Stanford study uses US payroll data ^[5], and several of the other figures cited here also come from US data. Results for other economies may differ.

What data comes next?

The sources give no release schedule. The Stanford paper was first documented in August 2025 and revised in August 2026 ^[5], so further updates are plausible. Earlier editions of the WEF report appeared in 2020 and 2023 ^[1]. Check the original publishers for newer figures before making a decision on them.