Generative AI and Automated Decision-Making Tools Usage Policy
Last Updated: September 2024
This policy provides guidance on the proper use of GenAI while working at the Good Food Institute APAC (GFI APAC). GenAI tools include OpenAI’s ChatGPT, DALL-E, Google’s Bard, and other similar applications that mimic human intelligence to generate work products or performance tasks.
GFI APAC is committed to adopting new technologies to aid our mission when possible and recognises that GenAI tools may offer attractive opportunities to streamline our work and increase efficiency. However, these tools also carry significant risk when it comes to privacy & security, accuracy, and intellectual property.
This policy discusses the unique issues raised by GenAI and provides guidelines for responsible use that are consistent with our organisational values, ethical standards, and regulatory requirements.
Considering GenAI is a rapidly evolving technology, this policy will periodically be updated to address new GenAI capabilities and changes to the regulatory framework.
Please reach out to apac-ops@gfi.org if you have any questions or comments about this policy.
Policy scope
This policy applies to all employees, interns, and contractors of GFI APAC who use GenAI tools as defined above.
GenAI common use cases
GenAI is a type of artificial intelligence that can create new content, such as text, images, audio, and video by learning from existing data. GenAI models can perform a wide variety of functions depending on the input and model type:*
| Model type | Applications | Common use cases |
|---|---|---|
| text-to-text | Text generation Classification Translation Summarisation Research Extraction Content editing | Email marketing content Note-taking Transcription Code generation Code documentation Sentiment analysis Question answering |
| text-to-image | Image generation Image editing | Design Media/Social |
| text-to-video | Video generation Video editing | Video production |
| text-to-task | Software agents Virtual assistants Automation | Helpdesk agent Robotic Process Automation (RPA) |
Guidelines for using GenAI tools
The use of GenAI tools is permissible while performing work for GFI APAC subject to compliance with this policy.
GenAI systems must be used in compliance with all applicable laws and regulations, including data protection, privacy, and anti-harassment and discrimination laws. These technologies must not be used to create content that is discriminatory, harassing, or otherwise violates the principles outlined in GFI APAC’s Code of Conduct and Anti-Harassment and Discrimination policies.
All AI-generated content must be reviewed for accuracy before relying on it for work purposes. If a reliable source cannot be found to verify factual information generated by the chatbot, that information cannot be used for work purposes.
Note that even with GFI APAC’s website disclaimer noting that we use AI in our work, published work products such as industry and research reports, or data visualisations, should specifically include attribution to AI if AI is used in the development of those products.
A summary of key guidelines:
- Refrain from inputting confidential information into GenAI Tools used without an enterprise subscription assigned by GFI APAC.
No confidential, proprietary, or sensitive data of any kind may be submitted (copied, typed, etc.) into publicly available GenAI third-party platforms accessed without an enterprise (organisation-wide) subscription, including but not limited to OpenAI’s ChatGPT, DALL-E, and Google’s Bard, regardless of the settings selected or the assurances made by its creators.
- Avoid inputting any personal information into GenAI Tools used without an enterprise subscription assigned by GFI APAC.
No personal data, including names, emails, addresses, IP addresses, phone numbers, photos, or likenesses, may be inputted.
- Review all generated content for compliance with the GFI APAC Code of Conduct and applicable anti-harassment and discrimination laws.
- Independently verify any output for accuracy.
Cross-check any AI-generated data outputs with additional sources of information.
- Disclose usage of any AI-generated externally facing content.
Note that if AI-generated content is initially produced for internal use but later adapted for external audiences, users must remember to disclose the AI use when the content is shared externally.
- Provide prominent notice prior to transcribing/recording meetings.
Note that before transcribing or recording materials, the meeting organiser must provide prominent notice to all attendees. Note that in certain scenarios, depending on the jurisdiction or anticipated end use, the meeting organiser may need a record of consent from the presenter and/or attendees. Consider whether a meeting requires transcription, and minimise the sharing of confidential or sensitive information in transcribed meetings.