Generative AI (artificial intelligence) is a type of technology that produces new content, including text, images, video, and audio. ChatGPT is one example of an AI chatbot, which can understand and generate human-like answers to text prompts.
It continues to evolve based on human feedback and interaction. Its main objective is to predict the next word in a sentence, based on the patterns of language contained within its training data.
It is the responsibility of all ACS team members to remain up to date on these guidelines and ensure they’re using AI appropriately and within proper ACS safety limits.
Fill out this form to request access.
Benefits vs. Risks
There are numerous benefits of using generative AI for certain tasks, like creating first drafts of emails or other business communication, conducting general searches to obtain background information, prioritizing lists or tasks, generating concepts and ideas, proofreading non-confidential documents, etc. However, there are also risks involved when using generative AI, as these applications may provide answers that sound correct and coherent but are factually wrong. Users may be unclear on whether they’re receiving information that is incorrect, plagiarized, or biased. Other risks include potential intellectual property infringement issues, breaches of confidentiality, data breaches, or other legal pitfalls.
To help mitigate these risks, ACS has created Guidelines and Guardrails, a set of rules that must be followed when using generative AI. These guidelines apply to all ACS activities. So, if you are aware of volunteers incorporating AI into their activities, please ensure that those projects are also reviewed for compliance.
Additionally, it’s vital for all team members to ensure that only approved meeting transcription or translation services are used during calls. Be mindful of meetings with external partners and participants, and ensure they fully understand that we do not allow any third-party AI bots in ACS-hosted meetings.
Requesting access
Generative AI certainly isn’t perfect, but there is no denying the rapid progress and future possibilities for its use. ACS is dedicated to ensuring team members use generative AI responsibly and adhere to the guidelines provided above.
Team members using any generative AI technology must request access using this form. In addition to providing your contact information, role within the organization, and a description of how you’re using generative AI, team members must read and accept the guidelines that have been approved by the Executive Team and Board of Directors and ensure that the terms of use for any generative AI platform do not contradict the guidelines. Expect a two-week turnaround time for a response to your request for access.
ChatGPT is blocked on ACS devices for anyone who hasn’t requested and received prior access. As generative AI technology continues to evolve, ACS reserves the right to review and block any site that could be a risk to the organization.
General rules
The following rules should be considered and followed when using generative AI.
- No confidential business information or sensitive personal information should be used in ChatGPT or other external, third-party generative AI environment.
- If generative AI content is being used internally as guidance and/or if that content will be shared outside of ACS, it must first be reviewed and approved by subject matter experts for accuracy and alignment with ACS policies. The source(s) for the generated content must be included in the review request to allow the reviewer to consider accuracy, bias, etc.
- If generative AI content is used in reports or presentations, it should be disclosed.
- While generative AI may be used as permitted by this guideline to increase efficiency, it should be viewed as a tool for assistance and should not replace personal responsibility, oversight, judgment, and analysis over one’s work.
- ACS team members should NOT use custom instructions while using their cancer.org email address to protect the organization.