Tonight's news about ChatGPT is not about launching a new model.
Nor is it about adding another feature.
Instead, its legal status in Europe has changed.
The European Commission announced today:
ChatGPT has been designated as a Very Large Online Search Engine under the Digital Services Act (DSA).
Abbreviated as:
VLOSE.
In Chinese, this translates to:
Super Large Online Search Engine.
Many people’s first reaction might be:
"Isn't ChatGPT just a chat AI?"
Why is the EU suddenly classifying it as a search engine?
This is exactly the key point to understand tonight.
The EU Does Not Say ChatGPT and Google Are Exactly the Same
Let's clear up this misunderstanding first.
VLOSE is a:
Legal classification.
It's not a:
product feature classification.
This does not mean the EU thinks:
ChatGPT is just another Google Search.
Rather, since ChatGPT can help users find, organize, and present information from the web, and its usage volume is already very high,
the "search for information" function starts to have social impacts similar to those of large search engines.
Reuters quoted the European Commission explaining:
ChatGPT meets the legal definition of an Online Search Engine under the DSA.
So the EU is no longer asking:
"Is this a chatbot?"
Instead, it asks:
"Has it become a major gateway for people to access information online?"
The answer is clearly yes.
How Large Is "Very Large"?
The DSA threshold is:
Over 45 million monthly active EU users.
About 10% of the EU population.
ChatGPT has not just crossed this line.
According to data OpenAI submitted for DSA compliance:
During the six months ending March 31, 2026,
ChatGPT Search averaged about:
159.1 million active recipients per month in the EU.
This figure is calculated by OpenAI specifically for DSA compliance purposes,
and should not be treated as a general monthly active user (MAU) commercial metric.
But regarding regulation, it clearly:
far exceeds the 45 million threshold.
So the true signal tonight is:
ChatGPT is now so large that it can no longer be regulated as just "an AI app."
Does Being Designated as VLOSE Mean ChatGPT Is Illegal?
No.
This is very important.
Today's decision is a:
Designation.
It is not:
a judgment of illegality.
Nor is it:
a penalty.
The EU's logic is that:
when a service reaches a certain scale,
even without proof of wrongdoing,
it must take on higher-level risk management responsibilities.
Like:
A small shop with ten employees
and an information gateway affecting hundreds of millions daily
should not be held to the same accountability standards.
So VLOSE essentially means:
Scale itself becomes a source of responsibility.
What Does ChatGPT Need to Do Next?
The EU states that after being designated as VLOSE,
the service must identify, analyze, and assess its own potential systemic risks.
That is:
Systemic risks.
This doesn't mean checking:
"Did it give a wrong answer yesterday?"
Instead, it asks:
If this service is used by tens or even hundreds of millions daily,
could recurring issues potentially affect society as a whole?
The EU highlights areas including:
Illegal content.
Fundamental rights.
Freedom of expression.
Media freedom and pluralism.
Discrimination.
Consumer protection.
Children’s rights.
Public safety.
Election processes.
Protection of minors.
Mental and physical well-being.
And public health.
Why Would AI Search Face These Issues?
Here's a simple example.
Previously, when searching:
"What are a certain candidate’s policies?"
Google usually gave you:
a list of links.
You still decide which to click.
Now when you ask AI,
it may directly:
organize,
compare,
and condense
several sources
to give you what looks like a summarized answer.
This is more convenient for users, of course.
But it also means:
AI is not just helping you find information; it’s deciding how to organize it.
Which sources appear first?
Which get omitted?
Which view is presented as the main conclusion?
If the response is wrong,
would users take it at face value?
If this only happens to one person,
it's a product quality issue.
If it happens daily to hundreds of millions,
it becomes a:
systemic risk.
Minors Are a Especially Important Concern
The EU is specifically including:
Protection of Minors
and
Mental and Physical Wellbeing
in the risk scope.
This aligns with ChatGPT’s recent product direction.
OpenAI launched ChatGPT for Teens in August,
using age information and predictions
to determine if a user is likely under 18,
and applying different safety settings accordingly.
This acknowledges even AI companies accept:
the same AI product cannot use identical logic for adults and teenagers.
Now the EU also legally requires:
Very large services to formally evaluate:
the potential impact of their design on minors.
Another Critical Issue: Elections
DSA requirements for VLOSE also include:
Electoral Processes.
This does not mean:
ChatGPT can no longer answer political questions.
The real question is:
If many people begin asking AI directly:
What did candidates say?
Is a piece of news true?
What impact does a policy have?
Or even:
"Who should I support?"
Then AI’s role in information distribution grows.
Misinformation,
manipulated sources,
biased summaries,
and large amounts of auto-generated false content
may all create challenges.
Regulators are no longer just concerned about:
AI hallucination errors.
They now ask:
What happens to society when a system serving hundreds of millions daily makes mistakes?
OpenAI Will Not Just Submit a Report and Be Done
Another key aspect of VLOSE is:
External supervision.
The EU’s VLOP/VLOSE rules require large services to:
set up internal compliance functions,
and undergo at least one independent audit per year.
They must provide relevant data to:
the European Commission,
and national authorities,
so they can monitor regulatory compliance.
Selected, reviewed researchers can also apply to access platform data under specific conditions,
to study systemic risks of these large services.
This is a major change.
In the past, AI companies mainly conducted risk assessments themselves,
publishing their own reports
and sometimes collaborating with external research groups.
In the future, within the EU,
the largest services will increasingly face:
institutionalized external oversight.
Does This Mean the EU Can See Your Private ChatGPT Conversations?
No.
"Regulators and researchers can access data"
does not mean:
any researcher can open someone’s private chat logs.
The DSA data access framework includes:
eligibility,
purpose,
procedure,
data protection,
and research use restrictions.
The EU requires that:
large platforms cannot always respond to public risk issues with:
"Trust us, we are managing it."
Regulators and qualified researchers must have ways to obtain sufficient data to verify:
if these risks actually exist,
and whether company measures are effective.
What Happens in Four Months?
The EU’s public announcement states:
ChatGPT, Reddit, and Roblox have four months after designation
to fulfill the additional obligations.
The currently given deadline is:
End of November 2026.
This does not mean that on December 1,
you will suddenly see ten new buttons when opening ChatGPT.
Many changes are likely to occur behind the scenes, such as:
internal risk assessments,
audits,
data retention,
researcher data access mechanisms,
compliance teams,
risk mitigation measures,
model and product design,
and content handling procedures.
If risk assessments find a feature causes systemic problems,
DSA allows requiring services to:
change design,
adjust system operations,
or modify recommendation and information presentation methods.
So in the long run,
the regulations may ultimately focus on:
how products are designed.
Are the DSA and EU AI Act the Same Law?
No.
This is also an easy confusion.
The EU AI Act mainly covers:
AI systems,
AI models,
high-risk uses,
transparency,
banned uses,
and general-purpose AI issues.
The DSA manages:
risks large online services pose to users and society.
So the same ChatGPT
may be subject to multiple legal frameworks.
You can simply think:
AI Act asks:
What rules should this AI itself follow?
DSA is more like asking:
When this service is large enough to impact society, what responsibilities must it bear?
They do not replace each other,
but may overlap.
Why Is Tonight’s Development More Important Than Typical AI Rules?
Because ChatGPT’s role is evolving.
Back in 2022,
it was mostly seen as:
a very advanced chatbot.
Later, it became:
a writing tool,
research tool,
coding tool,
a work assistant.
Now it is also:
searching the web,
organizing news,
comparing products,
looking up information,
and often the first info gateway for many people facing questions.
User habits are shifting from:
"Let me Google it first,"
to:
"Let me ask ChatGPT first."
Regulations naturally have to move accordingly.
This is why today’s VLOSE designation matters so much.
What Does This Mean for Ordinary Users?
In the short term:
ChatGPT will not suddenly become unusable because of tonight’s decision.
Nor does this mean the EU is banning generative AI.
What is likely to happen is:
The large AI services you use
will have to more formally explain:
how they manage risks,
how they protect children,
what happens when errors occur,
which content is restricted,
how users can appeal,
and the social impact of their algorithms.
Whether external researchers can truly verify their claims.
These were once considered:
"good practices" of AI companies.
But now for the largest services,
they are becoming:
legal responsibilities.
The Most Important Change to Remember
In AI’s first development phase,
the question was:
How strong is the model?
The second phase asked:
How much work can AI do for me?
Now a third question is emerging:
When AI has become the information gateway for hundreds of millions daily, who monitors its impact?
By formally adding ChatGPT to VLOSE, the EU is saying:
Not that ChatGPT is "just like Google,"
but that:
Chat AI tools have grown so large, they are now considered public information infrastructure.
The greater the capabilities,
the more users,
the more the responsibility cannot remain:
"It’s just an AI answer."
Today, progress together with AI.
Learn one AI skill a day.
Save a little time daily.
Increase your ability bit by bit.
SasaDaily, growing with you.
Recommended Reading
ChatGPT Is the Same AI, So Why Does It Now Guess Your Age First to Decide What It Can Talk About?
