The most interesting change from AI in the workplace may not be that computers can write faster or create summaries in seconds. The bigger change is that more people are using AI to perform tasks that were previously outside their roles.
A small business owner can draft contracts, analyze sales figures, or fix a webpage without directly asking for help from three different people. A marketer can read customer data. An operations staff member can create simple tools to reduce manual work. The boundaries between professions are becoming more flexible.
Recent research from OpenAI Economic Research, which analyzed over 1.5 million job-related messages from April to July 2026, found that some AI usage occurs across job fields. Some of these new activities are also repetitive and are starting to become a regular part of users' workflows. ([openai.com](https://openai.com/index/unlocking-new-ways-of-working/?utm_source=openai))
From “performing tasks” to “expanding roles”
Until now, discussions about AI often start with the question: what jobs can be replaced by AI? That question is important, but it is not complete. There is another equally important question: what can someone do when technical, time, or skill cost barriers are lowered?
AI does not automatically make someone a legal expert, financial analyst, or programmer. However, AI can help that person get past the initial stages that are usually the most obstructive: understanding terminology, structuring information, finding patterns, drafting, or transforming raw data into discussion materials.
Imagine AI as a “versatile coworker” capable of creating the first version of many things. It can help prepare analyses, but it may not fully understand the business context. It can write code, but it may not be aware of security implications. It can summarize contracts, but it does not replace legal judgment.
This is where the human role shifts. Not just as executors, but as individuals who set goals, provide context, review results, and make final decisions.
Why is this change important for non-technical workers?
AI makes certain basic skills more accessible. Reading data no longer always means mastering an entire analytics suite. Creating prototypes does not always start from months-long programming projects. Preparing presentations, reports, or working documents can be done faster from materials that were previously scattered in many places.
OpenAI also reports that the use of AI agents is beginning to shift from brief conversations to tasks that require longer periods, such as research, data analysis, document creation, and workflow automation. ([openai.com](https://openai.com/index/how-agents-are-transforming-work/?utm_source=openai))
However, easier access does not equate to automatically improved competence. Someone can produce a report that looks professional without truly understanding the quality of the data. This is where new risks emerge: work becomes faster to produce, but harder to evaluate if people only look at the final results.
The risks: responsibility may lag behind capability
When someone can do more with the help of AI, there is a possibility that organizations will consider all those capabilities equivalent to professional expertise. However, drafting and being responsible for the impact are two different things.
For example, AI can help create simple financial analyses. But investment decisions still require an understanding of assumptions, risks, and business conditions. AI can assist in writing code to process customer data. However, one must still check whether the code leaks sensitive information or produces incorrect conclusions.
Another risk is that “role expansion” turns into “increased burden” without clear boundaries. If AI enables someone to handle the work of three departments, the organization may continue to add responsibilities without increasing time, support, or compensation.
Therefore, the impact of AI is not only determined by the model's capabilities. It is also influenced by how companies define authority, review processes, work targets, and decision ownership.
What does this mean for us?
For workers, the ability to use AI should not be understood as the ability to write long prompts. What is more important is the ability to transform problems into verifiable workflows.
- Identify the parts you want help with. Is AI being used to find options, draft, process data, or execute actions?
- Differentiate between drafts and decisions. AI results can be initial materials, but decisions that impact money, customers, law, or reputation need human review.
- Preserve important context. Provide goals, constraints, examples of desired outcomes, and assessment criteria. AI works better when the problem is clear.
- Check sources and assumptions. Don’t just check grammar. Ask where the numbers come from, whether the data is complete, and what might have been overlooked.
- Build skills that complement AI. The ability to assess, communicate, understand customers, and make decisions will become increasingly important as draft production becomes cheaper.
Simple steps to get started
Choose one repetitive task that is safe enough to test for a week. For example, converting meeting notes into a task list, summarizing weekly reports, categorizing customer inquiries, or drafting internal documentation.
- Write down the process that is usually done manually.
- Determine which parts can be done by AI and which parts still need human review.
- Use non-sensitive data examples for initial trials.
- Compare the results with the old process: is it really faster, more accurate, or just looks neater?
- Note any errors that arise before expanding usage.
This approach helps us see AI as a tool for redesigning work, not just a text-generating machine.
AI does not erase expertise boundaries, but changes how we cross them
Recent developments show that AI can help people take on tasks from other fields and make them part of their daily work. This opens up opportunities for individuals and small businesses to move faster and try things that were previously too expensive or complicated.
However, the ability to try does not mean permission to ignore standards. The broader the roles assisted by AI, the more important context, verification, and accountability become.
The future of work may not only consist of humans working alongside machines. We may see more people with broader roles: not because they master all fields, but because AI helps them reach new areas more quickly. The challenge is to ensure that this expansion enhances work quality, not just adds to the workload.
Sources & further reading
- How workers are unlocking new ways of working — OpenAI Economic Research
- How AI is expanding what people do at work — OpenAI Economic Research
- How agents are transforming work — OpenAI Economic Research
- Will AI models achieve the ability to improve autonomously? — Associated Press
– Rio Yotto @rioyotto