The job post asks for prompt engineering, ChatGPT, and large language modeling, then spends most of its space on reporting, stakeholder updates, and vendor coordination. That is the new hiring puzzle: AI is often a signal inside an existing job, not proof the job has turned into an AI job. For learners deciding what to study next, the difference matters.
Indeed Hiring Lab described the labor market in January 2026 as subdued, with postings flat or declining in many occupations, even as jobs mentioning AI were growing across many knowledge work roles. That is a selective signal, not a blanket requirement. The smart move is not to paste AI onto every résumé line, but to find where AI language maps to real work.
The growth is real, but it is not evenly distributed
According to Indeed Hiring Lab, its AI Tracker reached 4.2% in December 2025. That number is useful because it keeps the hype in proportion: AI mentions are rising, but they are still a slice of the overall posting market. A worker who reads that as universal demand will waste time chasing generic credentials instead of role specific evidence.
edX tells the same story from another angle. Its August 26, 2025 analysis said job postings requiring AI skills had increased every month in 2025, and that employers required AI skills in three times more job postings than they did two years earlier. Growth like that deserves attention, but it does not erase the old screening basics: domain knowledge, communication, systems fluency, and proof that you can get work shipped.
The occupational split is the career lesson
Indeed Hiring Lab gives the cleanest warning against treating every job as an AI job. In January 2026, it reported that nearly 45% of data and analytics postings contained AI related terms, compared with about 15% in marketing and 9% in human resources. Same labor market, very different signal strength.
That split should shape your learning plan. If you are in data and analytics, AI literacy is moving closer to the center of the screening conversation, so projects that show model assisted analysis, evaluation, or workflow automation carry more weight. If you are in marketing or human resources, AI may still help you stand out, but the stronger résumé evidence is usually a workflow story: what changed, what got faster, what became measurable, and what judgment you still supplied.
Titles are getting wider, not clearer
Indeed Hiring Lab reported in October 2025 that AI related job descriptions do not always explain what employers mean. A majority, 52%, mentioned building new AI tools or directly using AI models, while 14% cited using AI in recruitment. Roughly a quarter gave little context for how AI would be used in the role.
That is why the title alone is a weak guide. An AI Engineer posting may mean model development, application integration, data pipeline work, or a regular engineering job with an AI feature bolted on. Indeed also found that tech, management, and creative roles appeared to drive core AI use and development, while service and healthcare roles applied AI mainly in hiring processes. For applicants, the question is less what the title says and more whether the posting describes building models, using models, or merely being screened by them.
Learn the tools, then attach them to a workflow
edX identified AI, AI agents, ChatGPT, prompt engineering, and large language modeling as the top five AI skills in job posting data. That list is a decent starting point, but it is not a curriculum by itself. A certificate that teaches vocabulary without a finished artifact is thin signal in a pickier market.
A better approach is to pair AI literacy with the work you already want to do. Analysts can build a small portfolio around data cleaning, summarization, validation, and decision notes. Marketers can show campaign research, content operations, testing plans, and brand guardrails. Human resources workers can focus on policy, employee support workflows, and careful review of AI assisted processes, especially because Indeed found some AI mentions relate to recruitment rather than the job itself.
The next signal to watch is whether vague AI mentions become workflow specific requirements. When postings say only AI, GenAI, or ChatGPT, treat that as a prompt to investigate, not a reason to panic enroll. When they describe evaluation, automation, model use, or measurable process change, that is where your learning time can convert into interview evidence.