> ## Content Index
> Fetch the complete content index at: https://www.theleftshift.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Lovable to Train AI Models on User Prompts and Code by Default, Unless Developers Opt Out
- URL: https://www.theleftshift.com/lovable-to-train-ai-models-on-user-prompts-and-code-by-default-unless-developers-opt-out/
- Published: 2026-08-07T09:58:34.000Z
- Updated: 2026-08-07T09:58:34.000Z
- Description: The change takes effect on September 9, though users can opt out through the platform's privacy settings before then.
- Author: The Left Shift Bureau
- Tags: AI Startups

AI app-building platform [Lovable](https://www.theleftshift.com/lovable-launches-cloud-and-ai-tools-to-simplify-full-stack-app-development/) has updated its data policy to allow the company to use customer content, including prompts, code, project files, uploaded images and files, generated outputs, and usage data from Free and Pro users, to train and improve its [AI models](https://www.theleftshift.com/meta-says-its-ai-models-achieved-gold-medal-level-performance-across-top-global-stem-olympiads/) by default. 

The change takes effect on **September 9**, though users can opt out through the platform's privacy settings before then.

There was no official announcement, or a press release, instead, the announcement was burried in documentation published on July 31st, 2026.

*"Lovable may use customer data from Free and Pro plans for AI model training unless you opt out,"* the company stated in a changelog. 

According to Lovable's updated documentation, developers who opt out before the effective date will prevent their eligible data from being used for AI training. 

Those who opt out after September 9 will have future data excluded, however, content collected before the opt-out may already have been used. 

The company says it will not use end-user data from apps built on Lovable, or users' account and billing information, for AI training.

The update brings Lovable in line with a growing number of AI companies that rely on customer interactions to improve their models while offering varying degrees of user control.

[OpenAI uses ChatGPT conversations](https://openai.com/policies/how-your-data-is-used-to-improve-model-performance/?ref=theleftshift.com) from consumer accounts to improve its models by default, although users can disable training through their settings. 

Google similarly allows human reviewers to evaluate some [Gemini conversations](https://www.techradar.com/ai-platforms-assistants/gemini/gemini-warned-me-humans-might-review-my-chats-but-turning-it-off-comes-with-a-surprising-downside?ref=theleftshift.com) to improve its AI systems, with users able to manage activity and privacy controls. 

GitHub gives developers the option to opt out of using prompts and suggestions to improve GitHub Copilot. [LinkedIn](https://www.theleftshift.com/linkedin-will-now-use-your-post-to-train-ai-models-by-default/), another company owned by Microsoft, has a similar "opt out" deature. 

Adobe, meanwhile, has taken the opposite approach, repeatedly stating that customer content stored in Creative Cloud is [not used](https://helpx.adobe.com/in/creative-cloud/apps/generative-ai/non-adobe-models-in-adobe-products.html?ref=theleftshift.com) to train its Firefly generative AI models unless users explicitly submit content through designated feedback or improvement programmes.