According to Indeed's AI Tracker, the share of US job postings mentioning artificial intelligence terms reached 4.2% in December 2025, the highest level since Indeed began tracking the metric. Nearly 45% of data and analytics roles now contain AI-related terms, compared with about 15% in marketing and 9% in human resources. The surge comes even as overall hiring activity remains subdued, with total job postings just 6% above pre-pandemic baseline levels.
Here's what's actually happening. Jobs with AI mentions are bucking the overall trend and growing across many knowledge work occupations, while hiring activity remains subdued and job postings are flat or declining in many occupations. After peaking in early 2022, AI hiring declined through 2023, then reversed following ChatGPT's introduction. Now companies are flooding requirements across roles, but not because they're building AI infrastructure. They're either chasing investor narratives or genuinely unsure which positions will need AI fluency in 12 months.
Why it matters for you:
- The AI premium is fake for most roles: When 4.2% of all job postings mention AI but JOLTS shows job openings down 966K year-over-year, companies are inflating requirements without creating new positions. If you're hiring for roles that mention AI but don't genuinely require machine learning expertise, you're competing for talent in an artificially inflated market. Strip the AI language unless the role involves actual model training, fine-tuning, or ML infrastructure—you'll get better candidates who aren't chasing buzzwords.
- Salary negotiations just got murkier: Workers who added "prompt engineering" or "AI tool proficiency" to their resumes now have data showing 4.2% of jobs mention these terms. That's ammunition for asking for raises tied to "AI skills"—even when those skills amount to using ChatGPT. When discussing compensation, separate genuine AI/ML engineering (Python, PyTorch, model architecture) from AI tool usage (Copilot, Claude, ChatGPT). The former commands 20-40% premiums; the latter is table stakes and doesn't justify additional compensation.
- Hiring cycles will extend: The rise in AI-mentioned postings while total job openings collapse means every AI-adjacent role now gets flooded with applications from candidates gaming keywords. Your time-to-fill for any position mentioning AI, machine learning, or automation will extend 2-3 weeks as you wade through resume spam. Either remove AI keywords entirely and screen for actual skills during interviews, or explicitly state "this role does not involve AI/ML development" to filter out buzzword chasers.
Source: Indeed, AI Tracker (December 2025)
Watch this: The concentration is telling—45% of data/analytics roles mention AI versus 9% of HR roles. This isn't broad adoption, it's sector panic. If every data analyst job requires "AI experience" but total data analyst openings are flat, companies are raising bars without raising headcount. Expect the first wave of "AI-skilled" hires to underwhelm by mid-2026 when employers realize ChatGPT proficiency doesn't equal analytical capability.
The contrarian play: While competitors stuff job descriptions with AI keywords and drown in unqualified applications, write brutally honest job postings that explicitly state whether the role involves actual AI/ML work or just using AI tools. "This role does not require machine learning expertise; we're looking for strong Python developers who are comfortable using AI assistants" will attract better candidates and cut your time-to-hire in half. The AI keyword arms race is creating a massive adverse selection problem—be the company that opts out.