The Artificial Intelligence Safety Forum (AISF) is a nonprofit, self-regulatory forum for AI (Artificial Intelligence) safety in products that use generative AI.
The AISF helps consumers and businesses make informed decisions about products that use generative AI. This is done primarily through assessing and assigning a safety rating to these products.
Generative AI is a type of artificial intelligence that can create new content, like text, images, and music, by learning from existing data. It’s the technology behind products that can generate original work from a simple prompt.
A product with generative AI is an application or service that uses artificial intelligence models to create new, original content. This includes (but is not limited to) LLM based text generators, image generators, video and animation generators, audio and music generators, and multimodal generators.
We assess and rate both digital and physical products that use generative AI.
Digital products include (but are not limited to) apps, websites, and software platforms that use generative AI to create text, images, video, or other media. For example, AI companions, content generators, and chatbots available on phones, desktops, and browsers.
Physical products integrate generative AI into hardware such as toys, robots, wearables, or smart devices. These use AI models to interact with users in the real world, often blending voice, visuals, and movement.
AISF Ratings provide a clear assessment of a generative AI product’s safety, with each grade reflecting its performance across key safety categories from harm mitigation to data privacy. These ratings help consumers and businesses make informed decisions about products that use generative AI.
AISF Ratings include: A (Excellent Safety), B (Good Safety), C (Acceptable Safety), D (Poor Safety), and F (Critically Unsafe). You can learn more about these ratings here.
No, the AISF Rating system is voluntary.
AISF Ratings are assigned to products using generative AI using the following rating process:
Raters assess the product using generative AI using the AISF Benchmark System.
Raters provide a recommended AISF Rating.
Recommended AISF Ratings are reviewed, confirmed, and summarised.
The confirmed AISF Rating is applied to the product using generative AI.
AISF raters identities are confidential to reduce the risk of outside influence. AISF raters are not allowed to have any connections to individuals, groups, or organisations in the AI industry.
Registration is required to have your product with generative AI rated. You will need to designate an AISF person from your organisation and have them complete the form here to apply. Once registration is approved, they will receive information on the rating process. AISF will generally respond to registration requests within five business days.
We not only assess and rate the safety of products with generative AI, but also provide advisory services for developers and organisations. This means that if a team wants to strengthen safeguards or improve trust in their product, we can work with them directly – even if no public rating is requested. Our goal is always collaboration: helping creators build safer, more responsible AI products from the start.
We also offer education, research, and advocacy.
If we discover a product poses an urgent risk to the public, we notify the developer and the public at the same time. This ensures people are aware right away and can make safe choices while fixes are in progress. We also inform platforms and regulators, and always offer to collaborate with developers to address the issues quickly and effectively.
Yes. If you discover an issue that may pose an immediate and serious risk to the public, we encourage you to report this using our online form. Please include as much detail as possible about what you found. Our team will review reports promptly, verify issues, and take appropriate action, including notifying developers, platforms, regulators, and the public when necessary.