The EU is simplifying artificial intelligence rules: what does this mean for businesses?
On 7 May 2026, Euroopa Komisjon announced an agreement aimed at simplifying the rules for companies that use or develop artificial intelligence.
Simply put: the European Union is not abandoning oversight of artificial intelligence, but is giving businesses more time and wants to make the rules easier to understand.
This is particularly important for small and medium-sized enterprises, start-ups, and those who are only beginning to use artificial intelligence in their products, services, or work processes.
What happened?
The European Union has adopted the Artificial Intelligence Act, or AI Act. It regulates how companies may use artificial intelligence, particularly when the technology may affect people and important decisions.
For example, when artificial intelligence helps:
- select job applicants;
- assess pupils or university students;
- make medical decisions;
- analyse biometric data;
- work in the security sector;
- be part of critical infrastructure;
- control devices or other products.
Such systems are considered more sensitive because an artificial intelligence error can affect a person’s life, work, health, education, or rights.
Initially, companies were expected to adapt to the new requirements quite quickly: prepare documents, assess risks, explain how artificial intelligence works, verify the data used, and ensure human oversight.
In practice, this proved too complicated for many businesses, especially small businesses that do not have large legal and technology departments.
What is changing now?
The European Union decided to simplify some of the rules and postpone the deadlines for their implementation.
This means that companies will have more time to understand the requirements and prepare for them without rushing.
For some high-risk artificial intelligence systems, the new requirements will take effect on 2 December 2027.
For artificial intelligence embedded in regulated products, such as devices, toys, or technical products, the deadline is 2 August 2028.
In other words, companies are given more time to prepare documents, review processes, and understand which requirements apply to their products or services.
Does this mean that artificial intelligence can now be used in any way whatsoever?
No.
This is a very important point.
Simplifying the rules does not mean that artificial intelligence can be used irresponsibly.
The European Union is simply acknowledging that businesses need clearer rules and realistic deadlines. If artificial intelligence can harm a person, violate their rights, or is used in a dangerous way, oversight will remain strict.
For example, the ban on artificial intelligence tools that create intimate images of a person without their consent will be specifically strengthened. Such applications are considered dangerous and unacceptable.
What does this mean for an ordinary company?
If a business uses artificial intelligence for simpler tasks—for example, creating texts and marketing materials, automating service descriptions and customer responses, or analysing information—then the strictest requirements will probably not apply to it.
However, the main responsibility remains.
A company should understand:
- where it uses artificial intelligence;
- what data is transmitted to artificial intelligence systems;
- whether artificial intelligence can influence a decision about a person;
- whether the customer is not being misled;
- whether the results produced by artificial intelligence are checked;
- whether human oversight is in place for important decisions.
For example, if a specialist or small business uses artificial intelligence to improve a service description, post, or advertisement, that is one situation.
However, if a company uses artificial intelligence to automatically decide whom to hire, to whom not to provide a service, or how to assess a person, the risk is much higher.
Why is this important for Estonian businesses?
Many companies in Estonia already use artificial intelligence, even if they do not always refer to it as a separate technology.
Artificial intelligence can help write texts, process queries, respond to customers, analyse data, create images, prepare documents, or improve advertising.
This is convenient for a small business because artificial intelligence saves time and helps it work faster.
But now, merely using artificial intelligence is not enough. It is important for companies to understand where the line lies between a useful tool and a risky decision-making system.
If artificial intelligence helps draft a listing, article, service description, or advertising copy, the risk is usually low.
When artificial intelligence starts making decisions instead of a person, much greater caution is required.
These articles may also be of interest
A small business's growth journey: how a specialist's work becomes a genuine business system
A practical story about how a small business can become more systematic, measurable, and less dependent solely on one person's workload.The chemistry that brought together 3 million followers: an interview with the creator of the Thoisoi channel
A story about content creation, experimentation, and how knowledge can reach a very large audience in the digital world.The start-up grant is still open: what does a new entrepreneur need to know before applying for up to 20 000 euros?
A useful article for entrepreneurs considering starting a new business, applying for support, and preparing a realistic business plan.
What could a business do right now?
Although the deadlines are being postponed, it is not worth leaving everything until the last minute.
Businesses should already calmly review how they use artificial intelligence.
A few simple questions are enough to get started:
Where do we use artificial intelligence?
For example, in texts, advertising, customer support, analytics, recruitment, or data processing.
Do we use personal data?
If names, contact details, customer documents, or other personal information are entered into artificial intelligence, particular caution is required.
Does artificial intelligence influence decisions about a person?
If so, the risk is higher.
Does a person check the output of artificial intelligence?
Artificial intelligence can make mistakes, invent facts, or provide inaccurate answers.
Does the customer understand that artificial intelligence is being used?
In some situations, transparency is very important for trust.
Good news for small businesses
The main advantage of the agreement is that the European Union seems to have listened to companies' concerns.
Small and medium-sized enterprises often cannot meet complex requirements as quickly as large corporations. They have fewer resources, fewer specialists, and less time.
Therefore, simplifying the rules may help businesses fear artificial intelligence less and use it more calmly and consciously.
However, this does not mean that security, customer data, and fairness can be ignored.
Simple summary
- The May 2026 agreement does not mean that the artificial intelligence rules are being abolished.
Rather, it is breathing space for businesses.
The European Union tells companies:
“We understand that the rules are complex. We will give you more time and make the requirements easier to understand. However, artificial intelligence must still be used safely, fairly, and transparently.”
For businesses, this means less confusion, more time to prepare, and the opportunity to adopt artificial intelligence without undue panic.
At the same time, businesses already need to understand where they use artificial intelligence, what data they process, and who is responsible for the outcome.
The main thing to remember
Artificial intelligence can and should be used if it helps a company work faster and better.
But if artificial intelligence affects people, their rights, work, education, safety, or access to services, it is treated much more strictly.
The new rules of Euroopa Liit should not frighten businesses. They should help businesses use artificial intelligence consciously and safely, and in a way that gives customers greater confidence in the business.
Source: European Commission, 07.05.2026.