3 Million Daily Messages: How ChatGPT Became America's Unofficial Salary Negotiator

The days of awkwardly asking a coworker what they earn may be numbered. A new report from OpenAI reveals that Americans are sending nearly 3 million messages per day to ChatGPT about wages, compensation, and earnings — making the AI chatbot one of the most-used salary research tools in the country. For job seekers navigating an opaque labor market, this shift represents a fundamental change in how pay information flows. At Metaintro, we track how technology is reshaping the way people find work, negotiate pay, and build careers — and this trend may be the most significant development in compensation transparency since pay transparency laws began sweeping across states in 2023.

Why Are Workers Turning to ChatGPT for Salary Information?

The answer comes down to what economists call "information frictions." According to OpenAI's research, the price of labor is often hard to find and difficult to interpret — especially for workers who are early in their careers, switching fields, or moving to a new city. Searching for salary data is costly in terms of time and effort. Asking about pay directly can be socially awkward or even professionally risky. And informal networks distribute wage information unequally, meaning workers without strong professional connections are left guessing.

ChatGPT eliminates several of these barriers at once. Rather than requiring a worker to search across multiple websites like Glassdoor, Payscale, or the Bureau of Labor Statistics, interpret scattered salary pages, or ask a socially risky question in the office, the AI model can synthesize wage information from across the internet and return a personalized benchmark in seconds. For a first-generation college graduate entering the workforce with no family members in their industry, or a mid-career professional considering a pivot into a new field, this kind of instant, judgment-free salary guidance can be transformative.

The sheer volume — 3 million messages per day — also suggests that traditional salary research tools are falling short. Job boards and compensation databases have existed for over a decade, yet workers are increasingly choosing an AI chatbot over these established platforms. The reason is likely conversational specificity: rather than browsing generic salary ranges, workers can ask highly specific questions like "What should a senior marketing manager with 8 years of experience expect to earn in Austin, Texas?" and receive a tailored response.

What Exactly Are Workers Asking ChatGPT About Pay?

OpenAI's report breaks down the types of wage-related queries into distinct categories, revealing that workers use ChatGPT for two primary purposes: translating pay into a usable benchmark, and understanding what a specific role, company, career path, or business idea might realistically pay.

Among labeled wage-benchmarking messages, the breakdown reveals:

• Pay calculation — 26% of all wage queries (converting hourly to annual, understanding take-home pay after taxes, comparing compensation packages)
• Specific role pay — 19% (asking what a particular job title earns nationally or regionally)
• Entrepreneurship income — 18% (estimating what a freelance business, side hustle, or startup might realistically earn)
• Specific role at a company — 11% (asking what a software engineer makes at Google, or a nurse earns at HCA Healthcare)
• Occupation or career questions — 11% (exploring salary trajectories, comparing career paths by earning potential)

The pay calculation category being the single largest segment is particularly telling. It suggests that millions of workers struggle with something as fundamental as understanding what their paycheck actually means — converting between hourly and annual figures, accounting for taxes, or comparing a salaried offer against a contract rate. These are the kinds of basic financial literacy questions that many workers feel uncomfortable asking a manager or HR representative, but can ask an AI without judgment.

Which Industries Are Over-Indexing on AI Wage Searches?

Not all workers are turning to ChatGPT for wage data at equal rates. OpenAI's data shows that occupation-related wage searches are concentrated in specific fields, and the pattern reveals something important about where pay transparency remains weakest in the American economy.

The industries where wage searches disproportionately over-index include:

• Arts, design, entertainment, sports, and media — Creative professionals often work in project-based or freelance arrangements where compensation varies wildly between gigs, clients, and markets
• Management — Senior professionals evaluating whether their total compensation (base, bonus, equity) is competitive against market benchmarks
• Healthcare — A sector where pay can differ dramatically between hospital systems, private practices, geographic regions, and specialties
• Transportation and logistics — Drivers, warehouse workers, and supply chain professionals navigating a mix of hourly, per-mile, and salaried structures
• Sales and business operations — Commission structures, on-target earnings, and variable pay make it nearly impossible to compare offers without deep analysis
• Computer and mathematical occupations — Tech workers comparing total compensation packages that include base salary, RSUs, signing bonuses, and refreshers across employers

The common thread across all these sectors is that pay is harder to benchmark, more negotiable, or more critical to career mobility decisions. In industries where a standard salary band is publicly posted and rarely deviates — think government jobs or unionized positions — workers have less need to ask an AI for help. But in fields where compensation is opaque, variable, or highly dependent on negotiation, ChatGPT is filling an information vacuum that traditional salary databases have not adequately addressed.

How Accurate Is ChatGPT on Salary Data?

With millions of workers relying on AI for compensation insights, the accuracy question is paramount. OpenAI has taken a notable step toward accountability by introducing WorkerBench, a new benchmark specifically designed to evaluate ChatGPT's performance on labor market tasks that matter to workers.

In its first iteration, WorkerBench evaluated GPT-5.4 against 2024 Occupational Employment and Wage Statistics (OEWS) median wages published by the Bureau of Labor Statistics at both the national occupation and metropolitan area levels. According to OpenAI, the results showed high coverage, small bias, and numeric estimates that fall very close to the official government benchmarks.

This is significant for several reasons. First, it represents one of the first times a major AI company has formally benchmarked its model's wage accuracy against authoritative government data. Second, it creates a public standard that can be tracked over time — if future GPT versions become less accurate on wage data, the benchmark will reveal it. Third, it signals that OpenAI views worker-facing utility as a core product metric, not just an incidental use case.

However, workers should approach AI-generated salary data with informed caution. ChatGPT synthesizes information from across the internet, which may include self-reported salary data that tends to skew higher than actual medians. The model also cannot account for hyper-local factors like a company's internal pay bands, upcoming budget cuts, or the specific leverage a candidate has based on competing offers. AI salary data works best as a starting point for research — not as a final number to anchor a negotiation.

What Does This Mean for Pay Transparency and the Job Market?

The rise of AI-powered salary research is happening alongside a historic expansion of pay transparency legislation across the United States. As of 2026, 17 states and multiple municipalities have active pay transparency laws requiring employers to disclose salary ranges in job postings, covering an estimated 60 million workers — roughly half the U.S. workforce. States including California, New York, Washington, and Colorado now mandate that employers publish actual expected compensation ranges, not broad placeholders.

Yet the 3 million daily messages to ChatGPT suggest that legislation alone is not solving the transparency problem. Even in states with strong pay transparency laws, research from Cornell University indicates that overly broad salary ranges — such as "$50,000 to $200,000" — can undercut the effectiveness of these laws. Workers know the range exists, but they still cannot determine where they should fall within it. That is precisely the kind of question ChatGPT can help answer by factoring in experience level, location, industry benchmarks, and company size.

As Metaintro CEO Lacey Kaelani told TestGorilla, the shift toward skills-based hiring and workforce planning is fundamentally changing how compensation is structured. When employers evaluate candidates based on demonstrated skills rather than credentials or job titles alone, it creates a more dynamic — but also more opaque — compensation landscape where workers need better tools to understand their market value.

The convergence of AI salary tools and pay transparency laws could accelerate a broader market correction. When workers arrive at job interviews armed with AI-generated salary benchmarks that closely match government data, employers face increasing pressure to offer competitive starting salaries rather than relying on information asymmetry to keep costs down. For the hiring process, this is a net positive — it reduces wasted time on misaligned salary expectations and leads to faster, more efficient hiring cycles.

How Should Job Seekers Use AI for Salary Research?

The OpenAI data reveals that workers are already using ChatGPT as a salary research tool at massive scale. But using it effectively requires understanding both its strengths and limitations. Here are practical strategies for job seekers looking to leverage AI in their compensation research:

• Be specific with your queries. The more context you provide — job title, years of experience, city, company size, industry — the more accurate the salary estimate will be. "What does a nurse make?" will produce a generic national average, while "What does an ICU nurse with 5 years of experience earn at a Level 1 trauma center in Denver?" will produce a meaningfully different and more useful answer.
• Cross-reference with official data. Use ChatGPT as a starting point, then verify against the Bureau of Labor Statistics OEWS data, Glassdoor salary reports, and industry-specific compensation surveys. The WorkerBench evaluation shows GPT-5.4 tracks closely with BLS data, but cross-referencing strengthens your negotiating position.
• Ask about total compensation, not just base salary. Many workers focus exclusively on base pay, but the most significant gaps often appear in benefits, equity, bonuses, and retirement contributions. Ask ChatGPT to break down total compensation for your target role, including typical benefits packages.
• Use it for negotiation preparation, not just research. Ask ChatGPT to role-play a salary negotiation conversation, generate counter-offer language, or identify which aspects of a compensation package are most negotiable in your industry. The 18% of queries in the entrepreneurship category suggest workers are also using AI to evaluate whether freelancing or starting a business could match or exceed their current salary.
• Understand the limitations. AI models draw from publicly available data, which may lag behind real-time market shifts. If your industry just experienced a wave of layoffs or a hiring boom, the AI's training data may not yet reflect the new salary landscape. Supplement AI research with conversations with recruiters, industry peers, and professional networks.

What Does This Shift Mean for Employers and Recruiters?

The implications extend far beyond individual job seekers. When 3 million salary-related messages are processed daily, the aggregate effect reshapes the entire labor market dynamic. Employers and recruiters need to understand what this trend means for their hiring strategies.

Candidates are arriving more informed than ever. The information asymmetry that employers have historically relied upon during salary negotiations is eroding rapidly. When a candidate can generate a detailed salary benchmark in under 30 seconds — one that closely mirrors BLS data — lowball offers become not just ineffective but counterproductive. Companies that consistently offer below-market compensation will find their time-to-fill metrics increasing as informed candidates reject or ghost their offers.

As Metaintro CEO Lacey Kaelani told Spiceworks, the AI recruiting "doom loop" — where companies rely on automated screening that filters out qualified candidates while workers game the system to pass AI filters — is a symptom of a broader misalignment between how employers evaluate talent and how workers present their value. AI-powered salary transparency adds another dimension to this dynamic: workers who understand their market value are less likely to accept roles that undervalue them, forcing employers to compete on compensation rather than information asymmetry.

For recruiters, the practical takeaway is straightforward: publish competitive salary ranges upfront, be prepared for candidates who arrive with specific data points, and shift the value proposition of your offer toward benefits, culture, growth opportunities, and flexibility — the factors that AI cannot easily benchmark.

The Bigger Picture: AI as a Leveler in the Labor Market

The 3 million daily messages represent more than a product usage metric — they represent a structural shift in labor market power dynamics. For decades, employers have held the advantage in compensation negotiations because they possess comprehensive data about pay ranges, market rates, and internal salary bands. Workers, by contrast, have relied on fragmented information from job boards, word-of-mouth, and the occasional Glassdoor review.

AI tools like ChatGPT are compressing that information gap at unprecedented speed. A worker in rural Arkansas now has access to the same caliber of salary intelligence that a well-connected professional in Manhattan obtains through their network. A first-generation immigrant navigating the American labor market for the first time can get compensation guidance that previously required an expensive career coach or a well-placed mentor.

This democratization of salary information has real economic consequences. When workers negotiate from a position of knowledge rather than uncertainty, average compensation tends to rise toward market equilibrium. Underpaying employers lose talent to competitors who offer fair market rates. And the industries where AI wage searches over-index the most — creative fields, healthcare, management, and tech — are precisely the sectors where closing the information gap could have the largest impact on worker earnings.

OpenAI's decision to formalize this with WorkerBench also sets a precedent for the AI industry. If wage accuracy becomes a competitive benchmark — the way language model benchmarks currently track reasoning, coding, and factual accuracy — then AI companies will be incentivized to improve their models' labor market utility. That competition would directly benefit workers, creating an upward cycle of better tools, better data, and better outcomes at the negotiating table.

People Also Asked

Q: Is ChatGPT accurate for salary research?

A: According to OpenAI's WorkerBench evaluation, GPT-5.4 performs well against official Bureau of Labor Statistics wage data, showing high coverage, small bias, and estimates that track closely with government benchmarks. However, AI-generated salary data works best as a starting point. Cross-reference with BLS OEWS data, Glassdoor, and industry-specific surveys for the strongest negotiating position.

Q: What are the most common salary questions people ask ChatGPT?

A: OpenAI's data shows pay calculation leads at 26% (converting hourly to annual, understanding take-home pay), followed by specific role pay at 19%, entrepreneurship income at 18%, company-specific role pay at 11%, and career or occupation questions at 11%. The high volume of pay calculation queries suggests many workers struggle with basic compensation math that they feel uncomfortable asking a coworker or HR representative about.

Q: How many states have pay transparency laws in 2026?

A: As of 2026, 17 states and multiple municipalities have active pay transparency laws requiring employers to disclose salary information, covering an estimated 60 million workers. States like California, New York, Washington, and Colorado require employers to publish actual expected compensation ranges in job postings. However, the 3 million daily ChatGPT wage queries suggest that legislation alone is not solving the transparency gap, particularly when posted ranges are overly broad.

Wondering what you should earn? Metaintro delivers daily job market intelligence, salary insights, and career opportunities straight to your inbox — so you never walk into a negotiation without the data you need. Sign up today and take control of your earning potential.