Grant individuals the right to opt out of being included in AI training data

regulations data-privacy research-policy privacy ethics-in-research ai-governance ai-policy ai-regulation

Share this with your representatives, journalists and friends

Key insights

Many company and platform quotes describe existing settings, toggles, or contractual commitments, while advocacy and regulator quotes argue that enforceable rights are still inconsistent or absent.
The supportive case combines two different kinds of interests: personal-data privacy and creator control over copyrighted or cultural works.
Transparency recurs across the debate: supporters treat it as necessary for recourse and accountability, while some critics say current opacity is one reason opt-out systems become unworkable.
Several objections come from creator and rights-holder groups that want stronger protection, arguing that opt-out can shift enforcement costs onto individuals rather than denying creators' interests.
Some participants support user control but disagree over the default and mechanism: simple opt-out, opt-in, standardized technical signals, voluntary tools, or collective and licensed participation.

Arguments for

Consent and privacy rights

Proponents say AI training should not depend on non-consensual collection or scraping of personal data. Individuals should be able to choose whether to participate, object to use of their data in training datasets or models, and withdraw or request deletion, with especially strong protection for children's data.

We need regulation guided by human rights, a universal recognition around leaving no space for data collection and AI development without consent.

Ramanjit Singh Chima · Jul 7, 2026 · full quote
Companies must immediately cease the practice of unlawful non-consensual web scraping of personal data for AI training purposes, and states must hold companies to account for their involvement in any human rights abuses linked to their design and business choices.

Amnesty International · May 28, 2026 · full quote
OpenAI must offer its users a choice about whether to participate or not in this practice.

Office of the Privacy Commissioner of Canada · May 6, 2026 · full quote
Individuals should have the right to opt-out, update, or request deletion of their data from generative AI tools. These rights should encompass AI training data and disclosure to other users of the tool.

American Medical Association · 2026 · full quote
Children and young people should have the right to opt out or withdraw consent for their data to be processed, and there must be adequate assurance that children’s data is kept for the shortest possible period. Due regard should also be given for the evolving capacity of a child,…

Alex Murphy · May 19, 2026 · full quote
right to object [...] Data subjects must be able to exercise their rights both: [...] on training datasets and, [...] on AI models

CNIL (Commission nationale de l'informatique et des libertés) · 2026 · full quote

Control over creative works

Supporters argue that creators, artists, writers, and performers should remain in control of whether their work is used to develop AI systems. They say AI training on creative works should require permission, and some add that fair compensation and separate, explicit AI-use consent are part of that control.

Our approach will ensure Australian writers, artists and journalists retain ownership over their work, meaning no company should use Australian creative works to train AI without the artist’s control.

Australian Government · Jul 15, 2026 · full quote
Creators and artists must remain in control of how their work is used to develop AI systems.

Universal Music Group · Jul 7, 2026 · full quote
It is vital that our government does not open the door to tech companies to train AI models on artists’ works without permission and pay.

Christian Zimmermann · Mar 24, 2026 · full quote
Creator content should not be used to train commercial AI models without their consent and fair compensation.

Darren O'Rourke · Feb 26, 2026 · full quote
This is simple: if your work is used to train AI, there should be a way for you, the copyright holder, to determine that it's been used by a training model, and you should get compensated if it was. We need to give America's musicians, artists, and creators a tool to find out whe…

Peter Welch · 2026 · full quote
Training AI on UGC. Layer 1 creator consent never covers AI training automatically. A separate, explicit AI use clause is required.

Mate Prgin · May 12, 2026 · full quote

Transparency enables enforcement

Proponents say opt-out rights are only meaningful if AI developers disclose what training inputs they used and whether rights-reservation signals were respected. Access to records and clear information is presented as necessary for accountability, legal recourse, and identifying unauthorized use.

Rightsholders told us that a lack of transparency from AI developers weakens their protections under copyright law by making it harder to identify infringement of their material. In addition, they cannot monitor whether any rights-reservation signals they may have used to prevent…

House of Lords Communications and Digital Committee · May 13, 2026 · full quote
Unfortunately, it is exceedingly difficult for copyright owners and artists to know if and how AI systems have used their work for training purposes. The TRAIN Act would provide a path to information sharing without the need to resort to infringement lawsuits.

Tracey Armstrong · Jan 22, 2026 · full quote
The TRAIN Act will give human creators the tools to peel back the curtain of AI model training and create a much-needed level of accountability for AI companies.

Emio Zizza · Jan 22, 2026 · full quote
By creating a pathway for artists and rightsholders to learn what inputs are used by AI programs, the TRAIN Act will increase transparency and allow effective recourse for creative works that have been used without consent.

Dr. Moiya McTier · Jan 22, 2026 · full quote
There is no path for creators to know if their work has been used — without their permission and without compensation — to train AI models. The TRAIN Act would grant artists access to such AI records.

Madeleine Dean · Jan 22, 2026 · full quote
Transparency efforts like the TRAIN Act mark a first, but absolutely critical, step towards ensuring that the production workforce continues to benefit from how their labor is used in the marketplace.

Matthew D. Loeb · Jan 22, 2026 · full quote

Practical user controls

Many services already provide toggles, settings, or terms that let users exclude past or future data from training, sometimes while continuing to use the product. Supporters present these practices as evidence that participation in AI training can be a user decision rather than a condition of service.

For users logged into a Microsoft account, you will be able to control whether your consumer data is used for training the generative AI models used in Copilot and other AI offerings from Microsoft through an opt-out control. Opting out will exclude your past, present, and future…

Microsoft · 2026 · full quote
Customer may, at any time, and on a go-forward basis, elect to opt-out of Docusign’s AI Training utilizing the applicable “opt-out” toggle or feature within the applicable Docusign Service.

DocuSign · Jul 1, 2026 · full quote
You can control whether HubSpot uses your account's customer data to train its AI models by using the AI model training toggle.

HubSpot · Jul 1, 2026 · full quote
We do not believe that users should be required to allow their submissions to be used for AI training simply in order to participate in contests.

AtCoder · Jun 29, 2026 · full quote
If you previously opted out of the setting allowing GitHub to collect this data for product improvements, your preference has been retained—your choice is preserved, and your data will not be used for training unless you opt in.

GitHub · Mar 25, 2026 · full quote
We've added a setting that lets you opt out of having your channel content used to train generative AI content models across Amazon.

Twitch · Aug 12, 2026 · full quote

Arguments against

Unworkable at scale

Critics say broad opt-out regimes are proving unworkable in practice and place an unfair burden on individual creators or rightsholders. At the scale of AI training, they argue, people often cannot tell whether their material was used or realistically manage exclusion themselves.

I am pleased that the Government’s response today confirms what it set out in March: that it no longer has a preference for introducing a broad copyright exception for AI training with an opt-out mechanism. It was clear that this approach would have been unworkable and placed an …

Baroness Keeley · May 15, 2026 · full quote
The current opt-out framework, which is meant to enable rights holders to protect their content and stop AI companies from using it for training if they say so, is turning out to be unworkable in practice.

Arthur Mensch · Mar 24, 2026 · full quote
Bectu has been clear that any model which places the burden on individual creators to ‘opt-out’ of the use of copyrighted work for AI training is unworkable, unfair and disproportionately harms freelancers.

Philippa Childs · Mar 18, 2026 · full quote
This proposal would have significantly weakened UK copyright law, with creators opted-in to AI training by default, removing the need for creator consent and placing an unrealistic burden on creators to opt-out.

Creators Rights Alliance · Mar 18, 2026 · full quote
To be clear: opting all of this music out of generative AI training is all but impossible. Despite this, there are legislators in various countries who are being convinced, thanks to lobbying from the most powerful and valuable AI companies, that providing a right to opt out of …

Ed Newton-Rex · 2024 · full quote
AI needs creators and the data they provide—but that data must be paid for, not stolen. Ironically, big tech relies on the strong IP regime as well; I am sure no one was more surprised than I was when Sam Altman at OpenAI noted, with irony, that DeepSeek had exploited its open-so…

Dame Caroline Dinenage · 2025 · full quote

Obstacles to AI development

Opponents argue that expanded opt-out rights, licensing duties, or onerous compliance requirements can restrict access to the data needed to train AI well. In their view, narrower datasets and harder access could slow development and favor a few players who already control the needed data.

Restricting access to training data, whether through mandatory compensation, expanded opt-out rights, or onerous compliance requirements, can have the effect of restricting AI development itself.

Information Technology and Innovation Foundation · Jun 24, 2026 · full quote
If you create, at the kind of scale we’re talking about, an obligation to license or a right to opt out of training, then you create huge obstacles to the capacity to access the data you need to train AI systems well. You create a system where there are limits in the datasets and…

Carys Craig · Feb 26, 2026 · full quote

Wrong mechanism for creators

Some argue that a new individual training-control right would not necessarily empower creative workers, because standard contracts could require them to sign that right away. Others say clearer answers lie in collective licensing, separate opt-in licensing, or voluntary tools rather than an individual opt-out right.

If we create a new right to control training, that will not give creative workers control over AI training... It will give our bosses control over AI training, because they will amend our standard contracts to demand that we sign away our training rights to them.

Cory Doctorow · Jun 23, 2026 · full quote
What creators need after this pause is a firm commitment to copyright and neighbouring rights and support for collective licensing for AI uses, including via existing trade union collective bargaining mechanisms.

Paul W Fleming · Mar 18, 2026 · full quote
This pioneering licence will enable rightsholders to separately opt-in to their content being used for AI model training and RAG models, providing a clear, legal framework for the use of copyrighted works in generative AI.

The Publishers Association · Mar 2026 · full quote
One way to reflect and respect those wishes is through existing and new voluntary “opt out” tools.

Business Software Alliance · Jan 28, 2026 · full quote

Middle ground

Prefer opt-in defaults

Several quotes support user control but say it should not depend on people discovering and reversing default inclusion after the fact. They favor the more privacy-protective default or participation only when users deliberately choose it.

The idea is that you have to build your systems to collect only what you need and nothing more. And, if one of the options is more privacy protective than the other one, then by default, the more privacy protective option needs to be pre-selected.

Woodrow Hartzog · Jul 16, 2026 · full quote
Developers should have the right to decide when and how their work enters the AI training pipeline. Some will want to participate, and that’s great. Others won’t, and that’s equally valid. The point is that it’s a decision, not a default.

Sean Roberts · Apr 13, 2026 · full quote
If you previously opted out of the setting allowing GitHub to collect this data for product improvements, your preference has been retained—your choice is preserved, and your data will not be used for training unless you opt in.

GitHub · Mar 25, 2026 · full quote
We do not believe that users should be required to allow their submissions to be used for AI training simply in order to participate in contests.

AtCoder · Jun 29, 2026 · full quote
The opt-out approach does not work in practice. Companies don't have the contract details of users and users don't know who is training based on their data. There is absolutely no public support for Meta or Google to include Europeans' personal data into their algorithms. Now all…

Max Schrems · 2026 · full quote

Standardized workable controls

Another strand supports exclusion rights but emphasizes that they must be technically viable, standardized, and documented. Machine-readable formats, publisher controls, and data-use records are presented as ways to make opt-outs effective while giving both rightsholders and AI providers legal certainty.

Rights holders should have the possibility to use tools enabling them to effectively exclude the use of their works in AI training in a limited number of standardised machine-readable formats, potentially managed and listed by the EUIPO, thereby enabling the effective exclusion o…

European Parliament · Mar 10, 2026 · full quote
By allowing rightsholders to opt out, the mechanisms must remain technically viable.

Computer & Communications Industry Association Europe · Jun 2026 · full quote
Model developers need to document the rights they have to work with the data they are using to train the model. This documentation should also provide information about the source of the data, whether it was public or private, etc. Model developers should respect the right of dat…

Ramayya Krishnan · Sep 12, 2023 · full quote
The publisher CR requires Google to provide publishers with effective controls over the use of their search content in generative AI.

Competition and Markets Authority · Jun 3, 2026 · full quote
We need clear rules for the use of copyright-protected content for AI training. Legal certainty would let AI developers know which content can be used and how licences can be obtained.

Axel Voss · Mar 24, 2026 · full quote

Across the quotes, support is broad for giving people and creators more say over whether personal data or creative works enter AI training, and many organizations describe concrete controls already in use. The sharpest disagreement is over whether a formal opt-out right is sufficient and workable: some see it as a basic protection, while others say it shifts the burden onto individuals, should be replaced by opt-in or licensing structures, or could limit access to training data. The debate therefore turns less on whether control matters than on default settings, transparency, enforcement, and implementation.

Cast your vote:

Results (127 votes):

Total (127 votes)

For 109 (86%) Abstain 0 (0%) Against 18 (14%)

For (108)

Twitch

I am this person! Add a sourced quote Report comment

votes For and says:

We've added a setting that lets you opt out of having your channel content used to train generative AI content models across Amazon.
AI Verified source (Aug 12, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
xAI
Artificial intelligence company developing the Grok family of frontier AI models.

I am this person! Add a sourced quote Report comment

votes For and says:

xAI never trains on your API inputs or outputs without your explicit permission.
AI Verified source (Jul 27, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Personal Data Protection Commission (Singapore)
Singapore’s national authority for administering and enforcing the Personal Data Protection Act (PDPA).

I am this person! Add a sourced quote Report comment

votes For and says:

Where User Data is used for GenAI model development, and exceptions to consent do not apply, organisations must obtain meaningful consent by providing AI-Specific Notifications. These notifications should, to the extent practicable, provide: [...] (d...
more AI Verified source (Jul 20, 2026)

added 19d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Flewnt
Perth-based rapper and advocate for protecting First Nations cultural property in AI.

I am this person! Add a sourced quote Report comment

votes For and says:

Just permission in general on what is entered and not entered into the models is definitely a part of ethical AI.
AI Verified source (Jul 18, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Perplexity Support
Help center team

I am this person! Add a sourced quote Report comment

votes For and says:

Some data may be used to train AI models and improve search quality, unless you opt out through Account Settings.
AI Verified source (Jul 16, 2026) 1 of 2

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Woodrow Hartzog
Law professor, Boston University

I am this person! Add a sourced quote Report comment

votes For and says:

The idea is that you have to build your systems to collect only what you need and nothing more. And, if one of the options is more privacy protective than the other one, then by default, the more privacy protective option needs to be pre-selected.
AI Verified source (Jul 16, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Ben Winters
Consumer Federation of America director

I am this person! Add a sourced quote Report comment

votes For and says:

It is the perfect recipe for something that needs federal government intervention. That's what legislatures and governments are there for: to protect people where they are unable to protect themselves and constrain companies from doing things that ar...
more AI Verified source (Jul 16, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Government of Australia
Government of Australia speaking through its delegate at the UN Global Dialogue on Artificial Intelligence Governance in July 2026.

I am this person! Add a sourced quote Report comment

votes For and says:

No company should use Australian creative works to train AI without the artist’s control.
AI Verified source (Jul 15, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Australian Government
National government of Australia; publisher of official digital and AI policy guidance.

I am this person! Add a sourced quote Report comment

votes For and says:

Our approach will ensure Australian writers, artists and journalists retain ownership over their work, meaning no company should use Australian creative works to train AI without the artist’s control.
AI Verified source (Jul 15, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
BigTime
Professional-services automation software company publishing a 2026 AI Use Policy.

I am this person! Add a sourced quote Report comment

votes For and says:

BigTime does not use Customer Data or identifiable AI Interactions data to train, fine-tune, retrain, or otherwise improve any AI, machine learning, generative AI, or large language models, whether operated by BigTime or by third parties.
AI Verified source (Jul 10, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Swarms
AI agent infrastructure company and publisher of the Swarms multi-agent framework.

I am this person! Add a sourced quote Report comment

votes For and says:

You can opt out of model improvement at any time, with no effect on price or features.
AI Verified source (Jul 7, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Anna Olsarenk
UN representative of the Global Forum for Media Development, a network supporting media organisations worldwide.

I am this person! Add a sourced quote Report comment

votes For and says:

AI systems are trained on journalistic content, which is scraped without consent, without compensation, without credit.
AI Verified source (Jul 7, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Ramanjit Singh Chima
Policy director and public-interest technologist at the Association for Progressive Communications; speaker at the UN Global Dialogue on AI Governance in July 2026.

I am this person! Add a sourced quote Report comment

votes For and says:

We need regulation guided by human rights, a universal recognition around leaving no space for data collection and AI development without consent.
AI Verified source (Jul 7, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Universal Music Group
Global music company, speaking through its public-affairs representative on creators’ rights and AI.

I am this person! Add a sourced quote Report comment

votes For and says:

Creators and artists must remain in control of how their work is used to develop AI systems.
AI Verified source (Jul 7, 2026)

added 3mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Google
Search and advertising company

I am this person! Add a sourced quote Report comment

votes For and says:

Like your Search Services History, your saved media is also used to develop and improve Google services and technologies, including AI models and safety measures.
AI Verified source (Jul 6, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Character Technologies, Inc. (Character.AI)
AI companion chatbot company

I am this person! Add a sourced quote Report comment

votes For and says:

You can opt out of your personal data being used to train and improve Character.ai’s AI models.
AI Verified source (Jul 1, 2026)

added 1mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Human Data Rights Coalition
Advocacy coalition focused on data ownership, consent, transparency, and rights in AI training.

I am this person! Add a sourced quote Report comment

votes For and says:

There is no single universal opt-out yet — which is exactly the enforceable right the Human Data Rights Coalition campaigns for.
AI Verified source (Jul 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
DocuSign
Electronic-signature and agreement-management software company.

I am this person! Add a sourced quote Report comment

votes For and says:

Customer may, at any time, and on a go-forward basis, elect to opt-out of Docusign’s AI Training utilizing the applicable “opt-out” toggle or feature within the applicable Docusign Service.
AI Verified source (Jul 1, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
HubSpot
Customer relationship management and marketing software company.

I am this person! Add a sourced quote Report comment

votes For and says:

You can control whether HubSpot uses your account's customer data to train its AI models by using the AI model training toggle.
AI Verified source (Jul 1, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
AtCoder
Japanese online competitive-programming platform and contest organizer.

I am this person! Add a sourced quote Report comment

votes For and says:

We do not believe that users should be required to allow their submissions to be used for AI training simply in order to participate in contests.
AI Verified source (Jun 29, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Marvin
AI-powered research repository service.

I am this person! Add a sourced quote Report comment

votes For and says:

Your data may be used to train the LLM models. If you prefer that it is not used, you can opt out while signing up for Marvin.
AI Verified source (Jun 23, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
PDQ
Software company whose official June 2026 AI data-and-safety FAQ says customer data is not used to train AI models.

I am this person! Add a sourced quote Report comment

votes For and says:

By default, PDQ and its AI Subprocessors do not use Customer Data to train any models. We specifically have contractual agreements in place with our AI Subprocessors that prohibit the use of Customer Data to train their models.
AI Verified source (Jun 5, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Mistral AI

I am this person! Add a sourced quote Report comment

votes For and says:

Input and output data are used by default to train our artificial intelligence models, unless you opt-out of training.
AI Verified source (Jun 5, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Computer & Communications Industry Association Europe
European technology-industry trade association.

I am this person! Add a sourced quote Report comment

votes For and says:

By allowing rightsholders to opt out, the mechanisms must remain technically viable.
AI Verified source (Jun 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cluesmith LLC
Company that provides the Shannon personal AI assistant.

I am this person! Add a sourced quote Report comment

votes For and says:

We will never use your personal content (emails, messages, calendar entries) to train AI models without your explicit opt-in consent.
AI Verified source (May 30, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
BetaNYC

I am this person! Add a sourced quote Report comment

votes For and says:

In selecting AI products to use internally, we will prefer tools that allow us to opt out of having our data used in the product’s training data.
AI Verified source (May 29, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Amnesty International
Human rights NGO

I am this person! Add a sourced quote Report comment

votes For and says:

Companies must immediately cease the practice of unlawful non-consensual web scraping of personal data for AI training purposes, and states must hold companies to account for their involvement in any human rights abuses linked to their design and bus...
more AI Verified source (May 28, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Loading more opinions...

Abstain (0)

Loading more opinions...

Against (18)

Kate Gilchrist
Head of Content Legal and Legal Operations at the Australian Broadcasting Corporation (ABC).

I am this person! Add a sourced quote Report comment

votes Against and says:

An opt-out system as proposed by AI companies places burdens and costs on rights holders, as it would require ongoing surveillance to see whether copyright material is being used. We cannot scour the internet and ensure we are opting out on all of th...
more AI Verified source (Oct 6, 2026)

added 9d ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Information Technology and Innovation Foundation
US innovation policy think tank

I am this person! Add a sourced quote Report comment

votes Against and says:

Restricting access to training data, whether through mandatory compensation, expanded opt-out rights, or onerous compliance requirements, can have the effect of restricting AI development itself.
AI Verified source (Jun 24, 2026)

added 2mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate
Cory Doctorow
Author and digital rights activist

I am this person! Add a sourced quote Report comment

votes Against and says:

If we create a new right to control training, that will not give creative workers control over AI training... It will give our bosses control over AI training, because they will amend our standard contracts to demand that we sign away our training ri...
more AI Verified source (Jun 23, 2026)

added 3mo ago

Choose a list of delegates

to vote as the majority of them.

Unless you vote directly.

Delegate

Loading more opinions...

More h/ai-governance votes

Ban AI impersonation of real individuals without their consent

Governments should fund open-source AI safety tools and red-teaming infrastructure as public goods.

Ban predictive policing

Mandate the CERN for AI to build safe superintelligence

Mandate third-party audits for major AI systems

Back