In this article
Talk about artificial intelligence often jumps to dramatic predictions. The more immediate story is quieter: AI is steadily changing the ordinary tasks that fill a working day, from writing emails to sorting invoices.
Key facts
- 52% of US employees used AI at work at least a few times a year in mid-2026, more than double the share three years earlier (Gallup).
- 15% use it daily, and 30% a few times a week or more.
- 88% of organizations use AI in at least one part of their business, according to Stanford’s 2026 AI Index.
- Gains are uneven: most productivity benefits so far are concentrated in a small group of leading firms.
Who is using AI at work
Gallup’s workforce surveys show steady growth. In the second quarter of 2026, 52 percent of US employees said they used AI in their role at least a few times a year, 30 percent a few times a week or more and 15 percent daily.
| How often US employees use AI at work | Share (Q2 2026) |
|---|---|
| At least a few times a year | 52% |
| A few times a week or more | 30% |
| Daily | 15% |
At the level of organizations, the Stanford AI Index 2026 reports that 88 percent use AI somewhere in their business, but it also finds that productivity gains are concentrated in a small number of leading companies. Using a tool and changing how work is done are not the same thing.
Where it is making a difference
Writing and research
Many office workers use AI assistants to draft emails, summarize long reports and prepare first versions of documents. The time saved can be real, but results need checking; our guide to what large language models can and cannot do explains why.
Sorting and searching
AI is good at classifying large volumes of information, such as customer messages, invoices or support tickets, so staff can spend more time on the cases that need judgment.
Healthcare
Software can flag scans for a specialist to review first, or draft clinical notes from a conversation. The most successful uses support skilled professionals rather than replacing their decisions.
Logistics and planning
Retailers and transport firms use AI to forecast demand and plan routes. Weather services now run AI forecasting models alongside traditional ones, as we describe in our article on the science behind weather forecasts.
What Large Language Models Can and Cannot Do
The risks
- Errors: confident but wrong answers can slip into reports and decisions.
- Data leaks: staff may paste confidential material into public tools.
- Scams: criminals use AI to write more convincing fake messages; our guide to recognizing phishing emails covers the warning signs.
- Uneven effects: some tasks and roles will change much more than others, and not every worker has access to training.
Skills that matter now
- Writing clear, specific instructions for AI tools.
- Checking output against reliable sources.
- Knowing your organization’s rules on which tools and data may be used.
- Judging when not to use AI at all, for example in sensitive personal or legal matters.
The bottom line
AI is becoming a normal part of many jobs, much as spreadsheets and email once did. The people who benefit most tend to be those who learn where it helps, where it fails and how to check its work. For a primer on the basics, see our beginner’s guide to AI.
Tasks most and least affected so far
| More affected | Less affected |
|---|---|
| Drafting routine documents and emails | Hands-on work such as plumbing, nursing care or construction |
| Summarizing information | Work built on personal trust and relationships |
| Basic coding and data analysis | Decisions carrying legal or ethical responsibility |
| Customer service for common questions | Complex, unusual cases needing judgment |
How employers are responding
Organizations are taking different approaches. Some provide approved AI tools with contracts that limit how company data is stored and used. Others restrict AI to specific tasks, such as drafting or summarizing, and require human review before anything is sent to customers. Many are writing policies that set out which data may be entered into which tools.
Training is the gap many surveys highlight. Workers who receive guidance on how to use AI well, and when not to use it, are more likely to report real time savings. Without that, tools often go unused or are used in ways that create risk.
Frequently asked questions
Will AI take my job?
For most people the nearer-term change is to parts of a job rather than the whole role. Tasks that are routine and text-based are changing fastest. Learning to use the tools well is one of the best ways to stay valuable.
Should I tell my employer if I use AI?
Follow your organization’s policy. Many employers now allow approved tools but restrict what data can be entered into them.
Which AI tools are safest for work?
Tools provided or approved by your employer, with business agreements that limit how your data is stored and used, are usually safer than free consumer versions.
