AI Tool
Disclaimer
You are responsible for all decisions made, advice given, actions taken, and failures to take action based on your use of the AI tool. This AI tool uses machine learning models that generate predictions based on patterns in data. Output generated by machine learning models is probabilistic and should be evaluated for accuracy on a case by case basis, including by employing human review of such output. Output generated by the AI tool does not constitute legal advice. You must read the White Paper to better understand the functionality and limitations of the AI tool before use. You are also required to read the entirety of the Terms of Service.
User Data
We use third party services (‘AI software providers’) that are not owned or controlled by us and who can collect personal data from the user of the AI tool. We are not responsible for the privacy practices or the content of such Third Party Services.
Accordingly, you should refrain from inputting user prompts that contain sensitive data, such as personal information and intellectual property. Additionally, you are encouraged to read the terms and conditions and privacy policies of the AI software provider here.
Example 1: Packaging symbols
Example 2: Source text
Example 3: Country of origin
Example 4: Safety testing requirements
How should the AI tool be used?
1. The purpose of the AI tool is to help you navigate various compliance related texts.
2. The value lies in its ability to provide somewhat coherent responses to natural language. In other words, you can write questions and receive an “answer” that might be easier to understand compared to the original source texts.
3. All generated outputs must be compared to the source text. You must not only take the selected items into consideration, but the complete source text.
What the AI tool is not
1. The AI tool does not think or have “opinions”. It cannot interpret or understand anything. It generates a response by “guessing” the next word based on mathematical probability. Think of it as a more advanced form of auto complete.
2. The AI tool is not a “truth engine”. It “blindly” generates a response based on the LLM prompt without any fact check. It can also make up false information, which is called hallucination (See LLM output: LLM hallucination).
How do you select sources?
We normally create the collections based on the original regulation or directive text. Here are two examples:
PART 134—COUNTRY OF ORIGIN MARKING
Sometimes we need to include more than one source text in the collection. Further, we may also include official guidance pages and documents in the collection.
Recommended process
Step 1: Write your question
Step 2: Test Narrow, Medium, and Broad to see how the output changes
Step 3: Compare the output to the source items – Does it match or can you find errors?
Step 4: Compare the output to the original source text – Does it match or can you find errors?
Data
1. The AI tool does not operate as an archive or file storage service. You are solely responsible for the backup of generated outputs and other safeguards appropriate for your needs.
2. The AI tool sends the user prompt to an external data processor. Read the Privacy Policy for more details.
Terminology
Source text: Regulation, directive, or other text
Collection: Set of items created based on the source text
Source items or Items: Subsection of the collection, which is based on the source text
LLM: Large Language Model
User prompt: The part of the prompt written by the user (normally a question)
LLM prompt: The complete text (including user prompt, selected source items and static prompt text) sent to the LLM
Output: The “answer” generated by the LLM based on the LLM prompt
Beta Version
This AI tools is work in progress until further notice. We are working to improve the quality and consistency of the outputs. Part of this is due to us optimizing the system, but also because the underlying technology is far from perfect. Such limitations are covered in this document.
Risks Assessment
Detailed information about potential risks and system limitations are detailed under Risk Assessment. It is important that you read this part to better understand the limitations of the AI tool.
Identified Errors
See the error log at the end of this document for more information.
Terms of Services
You can read the complete Terms of Service here.