LinkedIn is extracting user data for its AI models without explicit consent

LinkedIn is extracting user data for its AI models without explicit consent

Without explicit consent, LinkedIn uses users' data to train its AI models, forcing them to opt out if they don't want their data exploited.

Article by Thomas Brewster for Forbes – Translated by Lisa Deleforterie

One big new tech company is using user data to train AI models, without first informing them and putting them on by default. Like Meta and Grok (a chatbot developed by xAI), LinkedIn automatically integrates users into training its own AI models, as well as those of unidentified “affiliates.”

LinkedIn is owned by Microsoft, which also has a significant stake in ChatGPT developer OpenAI. That means the Redmond, Washington-based tech giant will also use data from the professional social network to train its AI models. After releasing this information, LinkedIn clarified that user data will not be used to train OpenAI’s core models, but will be shared with Microsoft for its OpenAI-based program.

according to LinkedIn: The AI ​​models that LinkedIn uses to power generative AI features may be trained by LinkedIn or another provider. For example, some of our models are provided by Microsoft’s Azure OpenAI service. »

Spokesman Greg Snapper said: We use OpenAI models provided by Microsoft's Azure AI service, as do other customers or users of this API. When we use these models, we do not pass data to OpenAI so they can train their own models. »

When LinkedIn trains generative AI models, we work to minimize personal data in the datasets used for training, including using privacy-enhancing techniques to mask or remove personal data from the training data set. The platform explained that this does not entail AI models that generate content » With data from the European Union, the European Economic Area or Switzerland. The European Economic Area is the European Economic Area, which includes the 27 member states of the European Union plus Iceland, Liechtenstein and Norway.

See also  40 years ago, ARPANET moved to TCP/IP

If you live in a country where LinkedIn has started using your data to train AI models, it's easy to turn this option off. Go to the Data Privacy section of Settings and turn off ” Use my data to train AI models for content creation

However, privacy activists remain concerned about LinkedIn’s decision to automatically enroll users in training for many of its AI models. Mariano Delli Santi, head of legal and policy at the Open Rights Group, a UK-based privacy organization, told Forbes: The opt-out model once again proves that it is woefully inadequate to protect our rights. It is unreasonable to ask the public to monitor and sue every online company that decides to use our data to train AI. Explicit consent is not only a legal requirement, it is a matter of common sense. »

The UK Privacy Authority urged: Take urgent action against LinkedIn and other companies that believe they can escape justice.


Read also: OpenAI Signs Deal With PwC to Sell ChatGPT to Enterprises

You May Also Like

About the Author: Octávio Florencio

"Evangelista zumbi. Pensador. Criador ávido. Fanático pela internet premiado. Fanático incurável pela web."

Leave a Reply

Your email address will not be published. Required fields are marked *