Recommended Free Tools
LinkedIn’s U.K. AI-training pause was temporary, not a permanent ban. On September 18, 2024, LinkedIn said it was not enabling generative-AI training on member data from the United Kingdom, European Economic Area, or Switzerland “until further notice.” The move followed questions from the U.K. Information Commissioner’s Office (ICO), which LinkedIn said it was engaging with. On November 3, 2025, LinkedIn said it began using some U.K. member data for content-generating AI model training again, while giving U.K. members an opt-out control.
The practical takeaway is current: review the Data for Generative AI Improvement or GenAI Model Improvement Setting in LinkedIn’s privacy controls. Turning it off limits specified future training use, but it does not delete your profile or posts, undo training already completed, or control every AI system LinkedIn operates.
What LinkedIn actually halted in September 2024
The 2024 announcement concerned training or improving content-generating generative-AI models—systems that can produce text such as suggested posts or messages. It did not establish that LinkedIn stopped every form of artificial intelligence or machine learning.
| Activity | What the 2024 pause covered | What it did not establish |
|---|---|---|
| Generative-AI model training | LinkedIn said it was not enabling this training on member data from the U.K., EEA and Switzerland “until further notice.” | A permanent prohibition or a finding that all processing was unlawful. |
| Security, safety and anti-abuse systems | Not identified as part of the pause. | That these systems were switched off. |
| Personalization and recommendations | Not identified as part of the pause. | That recommendation or ranking models stopped operating. |
| Using an AI feature | Different from training a model; the announcement did not say all AI features became unavailable. | That inference-time processing and model training are the same activity. |
| Training already completed | Not reversed by the announcement. | That previously trained models were unlearned. |
LinkedIn’s original explanation is available in its September 18, 2024 terms and privacy update. Its help documentation also distinguishes content-generating models from security, safety, anti-abuse and personalization systems (GenAI model-improvement control).
#1 Best Overall
Timeline: pause, regulatory response and resumption
| Date | Event |
|---|---|
| September 18, 2024 | LinkedIn announced updates to its User Agreement and Privacy Policy and said generative-AI training on member data from the U.K., EEA and Switzerland was not enabled until further notice. |
| September 20, 2024 | The ICO welcomed LinkedIn’s confirmation that it had suspended training involving U.K. users while the company engaged further with the regulator. |
| November 20, 2024 | LinkedIn’s updated User Agreement was scheduled to take effect. This was not the date of the AI-training suspension. |
| November 3, 2025 | LinkedIn said it began using some U.K. member data and content to train content-generating AI models, with an opt-out setting. |
The contemporary account of the ICO response is reported by The Hacker News. The ICO’s response was described as continued monitoring and engagement—not as a published penalty, permanent ban or final finding that LinkedIn had violated U.K. law.
Why the ICO raised concerns
The immediate issue was LinkedIn’s privacy and terms update, which clarified that member information could be used to develop LinkedIn and affiliated products, including generative-AI models. In regions where the feature was enabled, the control initially operated as an opt-out.
That approach raised ordinary data-protection questions:
Rank #2
- Was the notice clear enough for members to understand that their information could be used for model development?
- What lawful basis did LinkedIn rely on, and was the processing necessary and proportionate?
- Was the opt-out prominent, effective and available before relevant processing began?
- Were the data used minimized and safeguarded, and was a data-protection impact assessment required?
- Should U.K. users receive a different regional treatment?
Those are regulatory questions, not proof of a final legal conclusion. LinkedIn’s later privacy materials state the company’s own position on its legal basis; that statement should not be treated as an independent determination that the basis is valid.
What changed for U.K. members on November 3, 2025
LinkedIn’s later explanation says that, beginning November 3, 2025, it included U.K. members in a program using some data to train content-generating AI models. The described data may include:
- Profile details.
- Public content posted on LinkedIn.
- Information used to improve experiences and connect members with opportunities.
LinkedIn says private messages are not included in this described U.K. training use. Public availability does not make information non-personal: a profile can reveal a person’s name, employment history, location, education and other professional details protected by data-protection rules.
The current U.K. explanation and regional framework are in LinkedIn’s updated terms and data-use help page. Regional labels and availability can differ among the U.K., EEA, Switzerland and other jurisdictions.
How to opt out of the specified model training
LinkedIn’s labels and screen layout can change by account, product version and region. The usual path is:
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →- Open LinkedIn and select Settings & Privacy.
- Open Data privacy.
- Find Data for Generative AI Improvement or GenAI Model Improvement Setting.
- Turn the control off.
- Save or confirm the change if LinkedIn presents a confirmation button.
- Recheck the setting later, particularly if LinkedIn changes regional defaults or account controls.
LinkedIn says the control can cover LinkedIn and certain affiliates, including Microsoft, depending on region. Its help guidance says members can continue using LinkedIn’s generative-AI features after opting out.
What opting out does—and does not—do
| Opting out does | Opting out does not do |
|---|---|
| Limit future use of your LinkedIn data and content for the specified content-generating model-improvement purpose. | Delete your profile, posts or other LinkedIn content. |
| Give you a preference control for the covered training activity. | Reverse training that LinkedIn says has already taken place. |
| Leave LinkedIn’s generative-AI features available, according to LinkedIn. | Automatically govern security, safety, anti-abuse, personalization or every other machine-learning system. |
| Apply according to the regional scope LinkedIn assigns to the setting. | Guarantee that feedback submitted through an AI feature is excluded from training. |
LinkedIn separately warns that feedback supplied through AI features may be used for model training in some circumstances. Avoid entering confidential business, client or personal information into an AI prompt or feedback form merely because the main opt-out is disabled.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.When the main setting is not enough
For objections involving processing beyond content-generating model improvement, LinkedIn directs users to its Data Processing Objection Form. The company’s GenAI FAQs explain the distinction between content-generating AI, other machine-learning uses, feedback data and objection procedures.
An objection or privacy request is not a promise of model deletion. Once data has contributed to a trained model, a platform may not be able to identify or remove every influence without a separate unlearning process; LinkedIn’s stated control expressly does not undo completed training.
Best Value
What individual users and employers should do
For individual LinkedIn members
- Turn off the GenAI model-improvement setting if you do not want profile or public-post data used for that stated purpose.
- Do not assume the setting controls every LinkedIn AI system.
- Do not submit confidential material in AI prompts or feedback.
- Use LinkedIn’s objection process if your concern extends beyond the named generative-AI training activity.
- Keep a record of the setting and date if the decision matters for your professional or privacy records.
For employers and privacy teams
- Review whether employee profiles expose confidential projects, customers, locations or business relationships.
- Update social-media and acceptable-use policies to address public posts, AI prompts and feedback submissions.
- Map employees by jurisdiction because LinkedIn’s controls and notices are region-specific.
- Assess whether vendor contracts, data inventories and impact assessments cover AI model development and affiliated processing.
- Give staff a process for recording objections and escalating sensitive disclosures.
Why the story matters beyond LinkedIn
The episode illustrates a broader problem in AI governance: an opt-out can be easy to click but difficult to make fully meaningful after training has occurred. It also shows why headlines saying a platform “halted AI data processing” can mislead. The 2024 event concerned a defined generative-AI training use, not every AI function, and the policy later changed by region.
For users, the useful distinctions are between pausing and banning, opting out and deleting, public and non-personal, and training and inference. For organizations, the harder question is whether an employee’s public professional information can enter a model-development pipeline without conflicting with confidentiality, notice or governance obligations.
Frequently misunderstood points
- The ICO did not publicly announce a permanent LinkedIn ban. It welcomed the suspension and continued monitoring and engagement.
- The 2024 pause is not the current U.K. status. LinkedIn says U.K. processing began again on November 3, 2025 under an opt-out framework.
- Private messages are not part of the later U.K. description. LinkedIn says the described training use covers certain profile and public-content data instead.
- Turning the setting off is not erasure or unlearning. It does not affect training already completed.
- Public posts can still be personal data. Public visibility does not remove data-protection obligations.
The Bottom Line
LinkedIn paused the specified generative-AI training on U.K. member data in September 2024 after ICO concerns, but it was not a permanent prohibition. LinkedIn says the practice resumed on November 3, 2025 with a U.K. opt-out setting. Disable that setting for future covered training use, and use LinkedIn’s separate objection process for broader concerns—but do not expect either step to erase content or reverse completed model training.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
