OpenAI Signals maps ChatGPT work use across 165 activities and nearly 100 countries

A common activity framework enables comparison

OpenAI Economic Research has released a new Signals dataset describing how people use ChatGPT for work across 165 activity categories and nearly 100 countries. The quarterly data runs from the first quarter of 2025 through the third quarter of 2026. It is intended for researchers studying how AI use differs by place, task and time.

The categories use Intermediate Work Activities derived from O*NET, allowing ChatGPT messages to be compared with broader occupational and employment information. Examples include preparing instructional materials, creating marketing content, programming and analysing business data. A shared taxonomy is useful because job titles hide common activities performed across many occupations.

Work-related use is becoming more diverse

Across 74 countries observed in every quarter, the ten most common activities accounted for an average 51.1 per cent of classified work messages in the third quarter of 2026, down from 56.0 per cent in early 2025. OpenAI interprets this as work use becoming less concentrated, with countries developing different mixes of activities over time.

The dataset does not measure the whole labour market. It covers work-related messages from individual ChatGPT plans and excludes enterprise conversations and Codex use. About 43 per cent of the employment-weighted task mix in mapped United States occupations falls into other activity categories not individually reported, including physical, driving and health-record tasks.

Privacy protections shape what the data can show

OpenAI says researchers do not read individual conversations. Automated systems classify messages after identifying information is removed, results for small groups are withheld and statistical noise is added. Those protections limit re-identification risk but can make fine-grained comparisons less precise. Researchers should account for uncertainty rather than treating small changes as exact behavioural shifts.

Signals can inform questions about adoption, specialisation and access, but it cannot by itself show productivity, job loss or worker wellbeing. Usage reflects who chooses the product and which tasks are easy to express in chat. Strong analysis will combine the data with surveys, employment statistics and workplace studies while stating the gaps created by excluded plans and activities.

What teams should validate before production

For implementation, teams should record the exact model, software, dataset, connector and policy versions behind every result. Representative evaluation must include normal work, boundary cases, deliberate misuse and component failure, with outcomes separated by user role, data sensitivity and consequence. A strong average should never hide a serious failure in a critical path. Changes should pass the same regression set before promotion, and operators need a documented fallback that preserves evidence rather than silently broadening access or lowering quality.

Procurement and planning should account for the human work required to review exceptions, maintain controls and investigate incidents. Published performance or efficiency claims are useful starting points, but production confidence comes from evidence collected in the organisation’s own environment. A named owner, measurable acceptance threshold, cost model and tested shutdown path turn a promising release into a service that can be governed over time. Where decisions affect safety, rights, money or regulated records, accountable professional review remains essential.

Security and operational evidence

Security review should map every identity, credential, network destination, stored artefact and administrative access path involved in the workflow. Logs need enough context to reconstruct who initiated an action, which evidence was consulted, what the system proposed and what ultimately executed. Sensitive prompts and traces should not be retained merely because they are useful for debugging. Retention, redaction and access must follow the underlying data classification, with a practical process for correcting or removing material when policy or law requires it.

Operational monitoring should measure quality and service behaviour together. Latency, throughput and availability matter, but so do unsupported claims, missed constraints, unsafe tool calls and the rate at which people overturn recommendations. Teams should rehearse dependency outages and partial failures before unattended operation. If a model, connector or external service becomes unavailable, the workflow should fail predictably, preserve its state and tell the user what remains unfinished. Clear degradation is safer than an agent improvising around a missing control.

Transparency should scale with adoption

Organisations should explain the scope and limitations of the deployed capability to the people affected by it. Users need to know when automation is involved, how to challenge an outcome and where a human can intervene. That transparency improves feedback and prevents a pilot result from being interpreted as blanket approval for broader autonomy. Expansion should follow demonstrated evidence, not novelty or pressure to match a competitor announcement. The durable advantage comes from dependable integration, careful measurement and learning that survives the next model or product update.

Finally, teams should schedule a review after real usage rather than treating launch as the end of governance. The review should compare expected benefits with observed quality, cost, incidents and user behaviour, then retire workflows that do not justify their complexity. A service that works well can expand through the same evidence-based process. A service that fails should be narrowed or stopped without losing the records needed to understand why. This feedback loop is what turns experimentation into accountable operational capability.