As AI integrates into the modern workplace, major companies are documenting their attitude to the technology by developing AI principles. These principles aim to guide AI use within the company and create a common understanding among employees of how the company believes AI should be used.
Whether your organisation is already using AI daily or just getting started, it would benefit from having a well-formulated set of AI principles. This article shouldn’t be seen as legal advice, and we don’t offer AI consulting services — but we wanted to share our internal experiences and explain what AI principles are, how to develop them, and show the positive benefits you’re likely to see after a successful implementation.
the role of AI principles
In contrast to an official corporate AI policy, your AI principles should not be a set of rules to be strictly followed. Instead, they should be a collection of statements that describe your company’s AI philosophy. They should guide your employees, not give orders, and colleagues who read them should get an intuitive understanding of your company’s attitude to different AI tools and use cases.
For example, your principles might state that humans should always have oversight when AI tools are used to make decisions – but they probably won’t define the specific oversight process. Details like that can be ironed out in the AI policy, or in another internal steering document.
It’s impossible to set out a list of rules for every possible AI application. So your AI principles should make it easy for employees to assess potential AI use cases and figure out if they’re appropriate or not.
how to develop your AI principles
get inspiration from others
You don’t need to start from scratch. When thinking about your company’s own attitude to AI, why not get inspiration from the tech giants that actually use these tools? Companies like Google and Microsoft have made their AI principles public, so when you’re looking for basic concepts to include in your AI principles, they could be a good place to start. Our own set of AI principles could also be a source of inspiration. These companies may look very different to yours, but their commitment to ethical, human-focused AI should be shared by every company.
build a diverse team from across the company
The task of creating AI principles shouldn’t just be left to one person. Different departments will use AI tools in very different ways, and you need all those perspectives for your principles to be relevant to the company. Marketing is probably already using generative AI tools to help draft texts for social media posts or blog articles, while the finance department might be using AI-powered tools to detect anomalies in accounts. AI is a far-reaching technology, not a single tool, and you should get a diverse group of users to give input when making any AI-related decisions.
remember your existing values
In order to have an effect, your AI principles need to feel relevant to employees and become anchored within the organisation. So rather than developing something entirely new, look at the company’s existing principles. They may be described in your core values, business principles, brand guidelines or your mission and vision documents. Wherever they are, they’re likely to be great source material for your AI guidelines. For example, if collaboration is a key value for your company, make sure you emphasise the importance of collaboration when using AI tools – for example, by collaboratively reviewing and validating decisions made by AI tools before they’re actually implemented.
revisit and adjust your principles over time
As the organisation begins to use AI in new ways and employees get more confident and competent at using AI to enhance their capabilities, you may want to evaluate and tweak your principles – possibly to cover new use cases that have arisen since you first published them. Additionally, revisiting your principles at regular intervals helps bring them alive and engages employees, preventing them from becoming yet another document that is easy to forget about.
the effects of creating AI principles for your company
they help employees become independent, confident AI users
Our own figures show workers are generally positive about the impact of AI on their jobs. But other studies, such as this report from the OECD, have shown a small but significant amount of concern from workers over AI’s impact on their long-term job security. By laying out a balanced, constructive attitude in your AI principles, you can remove these fears, and by defining the basics of what AI should and shouldn’t be used for, you can inspire employees to confidently start experimenting and learning by themselves. If designed and implemented correctly, your AI principles can create an environment where the positive effects of AI can flourish.
you’ll maximise the benefits of AI, and minimise the drawbacks
Generative AI tools and the latest generation of AI-powered software can boost creativity, productivity and efficiency. Workers who already use these tools can confirm their usefulness for streamlining repetitive tasks, understanding large amounts of complex data or helping humans overcome creative blocks. Your company can seriously benefit from AI, but only if it’s used properly. When employees understand and accept your AI principles, they’ll start seeing the effects of these advantages – while steering clear of potentially negative AI applications that can hold your company back, rather than drive it forward.
future-proof your organisation for AI developments
Your AI principles should be a description of your company’s general approach to AI tools, and not a list of specific rules – so even as AI technology advances and develops over time, your principles should stay relevant. That doesn’t mean you never have to make changes, but even as new, powerful tools are released, basic concepts like transparency, oversight and ethical behaviour will still be relevant.
start designing your own AI principles
Every company can benefit from creating its own AI principles. To help you get started, we’ve created an AI principles template, modelled on our own set of principles. It’s always challenging to start from scratch, so you’ll benefit from using the template as a jumping-off point in your own internal work. Feel free to add to or adapt our suggested points and turn the template into something your entire organisation can stand behind as it starts exploring AI.