How to craft a generative AI policy (+ free template)

Generative AI, or GenAI, offers great tools for increasing personal and organisational productivity and fostering innovation. However, it also comes with many new risks, which may result in bad customer experience, reputational damage, legal consequences and other unintended outcomes.

These uncertainties have led some organisations to adopt a so-called "stop policy", which prohibits the usage of GenAI altogether. This might seem like an easy solution, but it also limits their competitiveness and productivity in the long term. Also, it may be unfair for employees who are left behind in technical development. Instead, we recommend creating a GenAI policy.

By reading this article, you’ll learn:

What is a GenAI policy
What a good GenAI policy should include
How to craft a GenAI policy within your organisation

Plus, at the end of the article, you can download our free template for getting started with your own GenAI policy.

What is a GenAI policy?

GenAI policy is a clear guideline for responsibly using and governing the use of generative AI within your organisation. Generative AI policy should provide clarity on what is responsible conduct for everyone using GenAI tools or building GenAI-powered apps.

What should a good GenAI policy include?

A good GenAI policy always starts from your vision of where generative AI can add value and open new opportunities for your organisation. Building on this vision, the policy should cover the topics that are relevant and necessary for its intended audience — varying from all employees to developers building new GenAI apps. Make sure it is written in a way that stakeholders within your organisation can understand.

Here are 12 good practice themes that you can selectively or comprehensively apply to your GenAI policy, based on your situation and needs:

What you can and cannot do with generative AI
Following the rules and using generative AI ethically
Learning about generative AI
Keeping your data safe
Having people check AI’s work
Letting others know when you use generative AI
What to do if something goes wrong
Evaluating generative AI tools of third parties
Understanding what pre-trained model can and cannot do
Managing risks of generative AI tools
Collecting feedback on the outcomes of your generative AI

How to craft a GenAI policy within your organisation

To create your own policy for generative AI, it is important to discuss the themes within your organisation.

We recommend hosting workshops, engaging people from different parts of the organisation to define your GenAI policy together. The more people involved, the merrier: You may want to involve key stakeholders from legal, technical, business and customer service teams, as well as top management. Our experience is that these sessions can potentially become an exciting learning experience for all.

When your policy starts to be clear, make sure that official decisions and implementation take place, just like with any other policy in your organisation. Remember that the field of generative AI evolves quickly, which means you’ll need to update your policy regularly.

Free template: Get started with crafting your own GenAI policy

To help you get started, we have put together a free template for crafting your Gen AI policy. The template includes questions on each of the 12 recommended themes aimed at helping you craft your own GenAI policy contents.

In addition, the template includes some example content on what the policy could be for each theme. Just fill in the blanks and rewrite the policy for your internal stakeholders in an understandable and relevant way.

Download the template by filling in this form. We hope you’ll find this and the other resources helpful in your GenAI journey. And we’d love to hear about your GenAI policy workshops, too!

‍

How to craft a generative AI policy (+ free template)

More Insights

Saidot Announces Collaboration with Microsoft to Support End-to-End AI Governance

Quick guide: What are the risks of AI hallucinations, and how to mitigate them?

Saidot Sends Six-Fingered Gloves to Altman, von der Leyen, Zuckerberg in Call-to-Action on AI Risks

Get started with responsible AI.