The Register: LinkedIn started harvesting people’s posts for training AI without asking for opt-in

Source URL: https://www.theregister.com/2024/09/19/linkedin_ai_data_access/
Source: The Register
Title: LinkedIn started harvesting people’s posts for training AI without asking for opt-in

Feedly Summary: You’ll have to opt out if you don’t like it – EU and a few others excepted
LinkedIn started harvesting user-generated content to train its AI without asking for permission, angering netizens.…

AI Summary and Description: Yes

**Summary:** LinkedIn’s controversial decision to automatically collect user-generated content for training its generative AI without explicit consent has sparked significant backlash among users. The updated privacy policy and FAQ highlight the implications of this practice, particularly regarding personal data usage and user opt-out mechanisms.

**Detailed Description:**
LinkedIn’s recent actions have raised critical concerns regarding user privacy and data security, particularly in the realm of AI and generative technologies. The platform has begun to harvest user-generated content, including posts and articles, to enhance its AI capabilities without obtaining clear consent, thereby igniting user outrage over perceived breaches of trust. Here are the major points outlined in the update:

– **Privacy Policy Update:** LinkedIn has updated its privacy policy to clarify how user data, including personal posts, will be used to train its generative AI models.
– **Automatic Data Collection:** There are indications that this data collection commenced before the latest communications, suggesting that LinkedIn has been gathering content without user awareness.
– **User Opt-Out:** The platform provides a mechanism for users to opt out of data collection for AI training. However, the default setting is “On,” requiring users to explicitly disable it to prevent their data from being harvested.
– **Geographical Exemption:** Notably, users located in the EU, Iceland, Norway, Liechtenstein, and Switzerland will not have their data used for AI training, which suggests a compliance strategy aligned with stricter data protection regulations in those regions.
– **Handling of Personal Data:** LinkedIn has stated its intention to minimize the use of personal data in its AI training sets, mentioning the use of privacy-enhancing technologies to redact sensitive information.
– **User Sentiment:** The backlash is significant, with many users expressing their anger and describing LinkedIn’s actions as a violation of trust. Many have shared advice on how to opt out of the data collection process.

This situation highlights the ongoing tension between innovation in AI and the ethical considerations surrounding user privacy.

**Implications for Security and Compliance Professionals:**
– **Importance of Consent:** This incident underscores the necessity for transparent practices regarding user consent, particularly in data-intensive domains like AI.
– **Compliance with Regulations:** The divergence in data handling practices between users in the EU and those elsewhere highlights the importance of staying informed about different regulatory environments and ensuring compliance accordingly.
– **Privacy Enhancements:** The mention of privacy-enhancing technologies points to potential best practices that organizations should consider adopting to protect user data while utilizing AI capabilities.
– **Public Perception:** Companies must be wary of public sentiment regarding data privacy, as backlash can lead to reputational damage, increased regulatory scrutiny, and potential litigation. This incident serves as a cautionary tale for other organizations harnessing user data for AI initiatives.