← Back to the full issueComing soon visits
Choose

Your data to train AI: will you still be able to say no?

Listen to the analysis

AIVoices and adaptation generated with NotebookLM; video illustrations generated with AI. Targeted local cuts or cards assisted by AI where indicated. Automatic transcription may contain errors. Historical analysis of issue 007, without a current legal update. Later source captures do not establish the historical state. Original retained; no new voice.

Download captions

Read the transcript
This audio adaptation was generated with Notebook LM.
Welcome to our deep dive into the October 2026 edition
of the Law Right newsletter.
So picture this, you log in and realize a platform
you use every day is just, you know,
quietly feeding your private posts
into this massive AI brain.
- Yeah, which is a pretty terrifying thought.
- Right, and then suddenly this alarming headline pops up.
It claims your legal right to say no
is about to be completely erased.
- Which naturally makes people wanna panic, right?
Like just hit delete.
- Exactly, do you delete your account?
Or is there, you know, a better way to respond?
Today we are navigating the highly complex rules
of using your personal data to train AI.
- It is incredibly easy to see a headline like that
and just assume the battle for your privacy
has already been lost, but I mean,
panic usually leads to the wrong moves.
- Yeah, so let's unpack this alarm.
If this leak draft is just like advocacy noise,
what is the actual signal here?
- Right.
- If a listener is worried about their data today,
what real world rules actually protect them right now?
- Well, let's start with the panic itself.
On September 21st, 2026,
the privacy advocacy group NYOB
published a very strong criticism.
- Okay.
- It was criticism of a
council presidency compromise draft,
specifically draft number 1253526,
dated September 3.
What's fascinating here is that PDF published by NYOB
is purely a negotiation document.
- Wait, hold on, let me jump in.
So a draft is really just a proposal.
- Exactly.
Their statement is just an advocacy position
on ongoing talks.
It has not adopted legislation
and it is certainly not a court ruling.
- Oh, I see.
That means we shouldn't act like our legal productions
have already vanished just because, you know,
some politicians are having a brainstorm in Brussels.
- Precisely.
The current debate in Brussels has to be separated
from the laws that are actually governing your data today.
We just can't mistake policy advocacy
for a sudden change in your legal rights.
- Which means we need to look at the current reality
of the GDPR.
How does objecting actually work right now?
Because I think most of us just assume
clicking no works the exact same way everywhere.
- It definitely doesn't.
Under article 21 of the GDPR,
you really have to distinguish
between two types of objections.
- Okay, how so?
- Well, if a company uses your data
for direct marketing and you object,
that processing stops automatically, boom, done.
- Right, so that one's just binary.
- Exactly, but objecting to your data being used
for AI training is usually based
on what the law calls your particular situation.
- Particular situation, that sounds a bit subjective.
- It is legally distinct and it actually triggers
a balancing exercise between your rights
and the company's legitimate interests.
- Well, hold on, a balancing exercise.
Who gets to decide if a tech giant's legitimate interest
outweighs my privacy?
'Cause I feel like I know who usually wins that fight.
- Yeah, I get that.
But it's not a blank check for tech companies, thankfully.
If we connect this to the bigger picture,
we have to look at a recent EU data protection document.
- Right.
- Specifically, EDPB opinion 28/2024.
This opinion demands a highly concrete assessment
of legitimate interests.
- So they can't just make something up?
- No, they can't just wave a hand.
They have to actually prove the necessity
of using your data and pass that balancing test.
- Here's where it gets really interesting, though.
What if my data is just sitting publicly
on a public forum?
- Like an open profile.
- Yeah, does the company just get a free pass to scrape it
because I didn't hide it behind a private firewall?
- Not at all.
Public data is not unrestricted material.
Just because your data is publicly accessible online,
absolutely does not remove the conditions of the GDPR.
- Oh, wow.
Okay.
- EDPB 28/2024 makes it very clear
that an AI model trained with personal data
is not automatically anonymous.
- So practically speaking,
if deleting my account doesn't actually remove
my pass post from an AI's training data,
what should you, the listener, actually do?
- The Law Right newsletter offers very specific advice here.
First, you need to understand that a request for erasure,
like deleting your account,
is entirely different from an objection.
- Because deleting it doesn't prove the AI model
was retrained to forget you, right?
- Exactly.
Think of it this way.
Deleting your account is like taking your cake
out of the bakery window.
- Okay.
- But an objection is stopping them
from using your secret recipe in the first place.
AI models don't automatically unlearn data
just because the source file is deleted.
- Uh-huh.
Okay, so pulling the cake from the window
doesn't delete the recipe from their files.
What's the actual right move then?
- First, identify the relevant privacy notice.
You wanna find their stated purpose and legal basis.
Then make a precisely targeted request
objecting to that specific AI training.
- Instead of just hitting delete account
and hoping for the best.
- Exactly.
And make sure to keep dated copies of your request
and their response.
- So let me get this straight.
When I explain my particular situation for the subjection,
should I be sending them evidence
because it feels like a hilarious and terrifying Catch-22
if I have to send sensitive documents
just to prove my identity?
- It is a Catch-22.
- Right, you'd basically just be handing them
more data to train on.
- Which is exactly why the newsletter explicitly warns
against needlessly supplying new sensitive documents
just to prove your point.
You do not wanna accidentally give them more personal data
in your attempt to protect it.
- Okay, so the core message for you today.
Take a breath, understand the specific processing
and current laws, and don't panic over proposed draft.
- Definitely.
- Your legal protections haven't vanished overnight.
- And looking ahead,
it really raises a fascinating question.
We spend so much time talking about the mechanics
of opting out, but we rarely consider the downstream effects
on the technology itself.
- Which leaves you with a final provocative thought to ponder.
If our personal data ultimately shapes the minds
and behaviors of future AI,
how might the AI's understanding of humanity
become totally skewed if millions of users
successfully exercise their right to object?
- Yeah, by removing their unique voices
from the training pool.
- Exactly.
Do we inadvertently leave the future of AI
to be shaped only by those
who just don't care about their privacy?
It's definitely something to think about
the next time you see a headline
about your digital rights.

Europe is debating the conditions for using personal data. A proposal does not, by itself, remove existing rights.

Areas of lawPersonal dataArtificial intelligence
Keywords#GDPR#objection#training#EU proposal

Case statusSources revisited on 6 October 2026 · general information, subject to the particular circumstances.

Original conceptual illustration made with AI assistance; no real event or person depicted.
Image and licence. Original conceptual illustration made with AI assistance; no real event or person depicted.

01

What is being discussed in Brussels

On 21 September, noyb criticised a European draft addressing, among other matters, personal data and AI. The organisation published a Council Presidency compromise draft dated 3 September 2026. Its statement is an advocacy position on negotiations, not a court ruling or evidence that a reform has become applicable. The debate must be distinguished from the law currently governing a particular use. [2,3]

The practical question deserves attention now: under what conditions could a business reuse personal data for AI development, and how could an individual challenge it? The final wording, adoption and timetable must be checked before announcing a change in rights. A negotiating document does not settle those questions. [3]

02

Objection does not work identically in every case

The GDPR remains the starting point. Where processing relies on legitimate interests or a public-interest task, Article 21 provides for objection on grounds relating to the individual’s particular situation. The controller must stop unless it demonstrates the compelling legitimate grounds specified in the provision or needs relating to legal claims. Objection to direct marketing follows a separate rule, without that balancing exercise. [1]

This means neither that every training operation requires consent nor that legitimate interests automatically authorise any collection. Purpose, necessity and the interests involved require examination; special-category data raises additional conditions. In Opinion 28/2024, the European Data Protection Board stresses a case-specific assessment of models and processing. Publicly available data does not become unrestricted material for reuse merely because it can be read online. [1,4]

Erasure and objection are also different requests. Deleting an account does not prove that previous use has been reversed or a model retrained. A useful response explains what is processed, on what basis and what the request changes. Avoid treating an unverified technical assurance as proof that a particular use has ended. [1]

03

The Law Right approach: keep the request and the response

  1. Identify the relevant notice: controller, stated purpose, categories of data, legal basis and rights channel. Keep a dated copy. Do not supply new sensitive documents to explain an objection unless necessary. A targeted request helps establish the issue without needlessly expanding the information disclosed.
  2. State precisely what you are asking and retain the acknowledgement. Where Article 21 is relevant, explain the elements of your particular situation rather than assuming the processing is direct marketing. Have an incomplete response or refusal assessed against the actual processing and the competent authority’s remit.
  3. For businesses, map the information supplied to AI tools and the uses permitted by the contracts. A reform under discussion does not replace assessment of present compliance. The next step is a targeted check, not an announcement that all protection has vanished. Keep contractual promises separate from what the service actually does.
THE POWER TO CHOOSE · 02

Your data is not a detail.

Which data, which actor, which use? Start by making the processing visible.

Explanatory visualisation, not a measurement or an individual legal conclusion.
→

Before deciding yes or no, identify the processing, its legal basis and the right you are exercising.

Sources and verificationOpen the sources +

Watch the story

AIVoices and adaptation generated with NotebookLM; video illustrations generated with AI. Targeted local cuts or cards assisted by AI where indicated. Automatic transcription may contain errors. Historical analysis of issue 007, without a current legal update. Later source captures do not establish the historical state. Original retained; no new voice.

Download captions

Read the transcript
This video adaptation was generated with Notebook LM.
Welcome to This Explainer.
Today we're diving into the pretty complex reality
of your data rights when it comes to AI training.
We're looking at all of this strictly
from the perspective of October 7th, 2026,
pulling insights directly from the Law Right Newsletter,
edition 007.
Look, it is a truly chaotic time for digital rights right now.
So let's cut through all that noise
and look at the actual grounded legal facts.
Here's our agenda for today.
One, a fictional AI scenario.
Two, Brussels drafts an advocacy.
Three, how Article 21 works.
Four, assessing anonymity and interests.
And five, practical steps to take.
Okay, let's jump right into Section 1,
a fictional AI scenario distinguishing illustration
from reality.
So which data, which actor, and which use?
Just imagine for a second that a platform announces
it's going to use your posts to train a new AI model.
You log in, you find an object button,
and you click it.
Boom, done, right?
But then weeks later, headlines are screaming
that your right to say no is completely vanishing
under new laws.
What's the actual reality here?
Should you immediately fire off an angry letter
or maybe just delete your entire account?
Well, hold on a second.
Before you do anything, you really need to identify
exactly what processing is proposed
and what specific rules govern it.
Now, that's an area we just talked about.
It's completely fictional.
It's not a real event, and it's not a measurement
of any individual legal conclusion.
But it perfectly highlights a really fundamental necessity.
You just can't make big decisions
about your data based on assumptions.
Moving on to Section 2, Brussels drafts an advocacy,
separating proposals from law.
If you've been following the news around late 2026,
you've probably seen the organization NYOB
strongly criticize a European draft
concerning personal data and AI.
But here's the thing.
We have to actively separate legislative noise
from actual binding law.
There's September 21st statement,
that is purely an advocacy position
on ongoing negotiations.
Remember, complaints contain allegations.
They are not court rulings, and they aren't proof
that a reform has actually taken effect.
The actual PDF document that sparked all this debate,
officially numbered 12535/26,
is a council presidency compromise draft
from September 3rd.
The absolutely crucial word there is draft.
It's just a negotiation document published by NYOB,
not adopted legislation.
An ongoing debate in Brussels does not, by itself,
magically erase the existing rights
governing your data today.
All right, Section 3, how Article 21 works,
the bedrock of current GDPR law.
So what exactly are those existing rights?
Well, the GDPR remains our absolute starting point here,
specifically Article 21.
It separates objections into two very distinct camps.
First, you have an objection based on your particular
situation under Article 6 grounds.
Think of this as a balancing test.
It's subject to exceptions, and the controller has to stop
unless they demonstrate compelling legitimate grounds.
Now, contrast that with a direct marketing objection.
With direct marketing, there is no balancing act.
The processing strictly and immediately stops.
They just don't work the same way.
And hey, this brings up a really massive point of confusion.
Erasure and objection are fundamentally different
legal requests.
Simply rage-deleting your account does not automatically prove
that previous data use has been reversed,
or that a massive AI model was somehow untrained.
An unverified technical assurance simply isn't proof.
You need a proper response detailing exactly
what your request actually changes.
This takes us to Section 4, assessing anonymity
and interests, the institutional interpretation
from the EDPB.
How are the big institutions actually navigating all this?
The European Data Protection Board is crystal clear
in opinion 28/2024.
Relying on legitimate interests requires a
concrete case-specific assessment.
It comes down to three incredibly strict criteria,
assessing the specific interest,
evaluating the necessity of the processing,
and doing that balancing test against user rights.
You literally can't just assume legitimate interests
automatically authorize a data grab.
Then, of course, we have the persistent myth of anonymity.
The EDPB provides a vital reality check here.
An AI model trained with personal data
does not automatically become anonymous just by its nature.
You need a concrete assessment to actually establish that.
Here's another super critical takeaway from the EDPB.
Public accessibility does not erase GDPR conditions.
Just because your data can be read online by anyone
does not mean it magically transforms
into unrestricted material for reuse,
totally exempt from data protection laws.
The rules still completely apply.
Finally, section five, practical steps to take,
the law-right approach.
Let's bring this all down to earth with some practical actions,
straight from the newsletter article.
Now, while this advice belongs to the author
and creates no universal legal duties,
it's an incredibly solid framework
for where things stand in October 2026.
Step one, identify the exact privacy notice.
Step two, state your request precisely
and keep a dated acknowledgement.
You want to explain your particular situation
rather than just assuming it's for direct marketing.
And step three, assess any incomplete responses
against the actual processing taking place.
There's also some really grounded advice here for businesses.
First, you've got to map the exact information
supplied to AI tools.
Second, carefully check what the actual service does
versus what the contractual promises claim.
And most importantly, businesses have to remember
that a reform currently under discussion
absolutely does not replace the immediate need
to comply with the law as it stands today.
So as we wrap up this explainer,
I want to leave you with a bit of a provocative question
to chew on.
Are you effectively separating legal facts
from advocacy noise in your own data decisions?
Untangling ongoing draft proposals
from the precise legal realities
governing your data right now
is literally the most powerful tool you have.
Thanks so much for joining me.
Keep your focus on the facts
and I will catch you in the next explainer.
Read the full issue

Law Right library

Return to every issue.

Each new issue expands the collection without replacing what came before.

Your network may not share an email address

Secure sign-in will become active once the app has been approved by the network. For this first issue, enter your email below. Your network choice is recorded only as the source; nothing is posted to your account.

LexiTalk to Lexi