Introducing AI: Bring your people with you

AI can be worrying for people at every level of an organisation. Some are afraid their jobs will change – or even vanish. 

Higher up the organisation, AI can seem both too big to ignore and too hard to deal with. Even in the c-suite, you often find anxiety. People feel they should understand it – but they may not. 

The best solutions are those we discuss in our videos – things like education, openness and governance. 

Create a positive AI culture. Focus on what you hope people will do with it, not what you don’t want them to do. You can let them experiment with low-risk AI to get familiar with it. And you can encourage and reward innovation, within appropriate boundaries. 

Ultimately, AI is just the latest technology, albeit with tremendous potential, that businesses have had to adapt to. If it’s a problem, it’s partly a business problem – and business people are good at solving those.

If you want a pro-AI message, our experience is it has to come from the top. And it needs to be supported by behaviour at the top. 

Because, when people are being encouraged to engage with something or do something, but they don’t see it happening further up the organisation, then they tend to take it less seriously and it tends to be less successful. 

Business leaders also need to be open and realistic with their people. For instance, it’s probably a bad idea to say exactly what you’re going to do with AI and where you’ll be with it in five years. No-one knows that. 

But there is a lot of power in explaining what you’re trying to do, why you’re trying to do it and how far you’ve got. That can make people feel engaged, and empowered. 

It’s usually a better way of encouraging and retaining talent than trying to suggest that somehow you have all the answers. 

To learn more about this topic, tune into our Bandwidth AI series. 

Implementing AI successfully involves people from many parts of the business. 

You need people who know about data, compliance, procurement, project management, risk, technology – and who just understand the organisation’s culture. And in some cases, you may need to include specialists on issues like ethics. 

Some businesses have a steering group figuring out the issues and determining policy, and reporting to the board. We’re seeing some develop AI assurance functions to assess the risk of deploying a particular AI tool. And businesses that use impact assessments are building them out for AI by asking questions like:  

• is data going to be processed by an AI tool? 
• will there be automated decision making? 
• will the provider have access to the data and what will they be allowed to do with it? 

Doing this at an early stage enables the business to assess and manage risk in a cross-disciplinary way. 

If AI will be important to your business, you need to educate your workforce – bringing them with you and getting them aligned as soon as possible. 

You can’t suddenly impose it on them in two years’ time when it’s urgent. They need to learn how to use AI effectively, and they can’t do that overnight. 

Some people with have a natural inclination to using the technology. And while these people who are good at it will say ‘great’, in some cases unfortunately you’ll find that people who don’t understand the technology will decide the AI is useless and then act as blockers. 

Which means that bringing the business up to speed – with people who either get it or are prepared to learn, or who don’t want to learn and therefore depart – inevitably takes some time. 

When your business decides to move forward with an AI system, a well-drafted AI policy can be invaluable. As well as providing essential guardrails this can be a good way to raise staff awareness and help ensure that your investment generates commercial benefits.  

In practice, any AI policy should have a clearly defined scope and explain which AI system has been approved for use, and for what purposes. Certain uses may be expressly prohibited and it is important to also be clear in this respect.  

The policy should set clear guidelines on how to use the AI technology, describe the relevant risks, and what steps to take to address these. 

As part of this, it can be helpful to explain: 

• requirements to attend AI training; 
• instructions on how and when to report an AI-related incident or policy breach;  
• the need to indicate if and to the extent that work has been produced or undertaken using AI; and 
• restrictions on inputting certain data and materials into AI systems, including sensitive and/or proprietary information. 

Where the system has been supplied by a third party, the contents of the AI policy may need align with the provider’s T&Cs. Practical requirements, for example, in respect of the use of AI watermarks (detectable only by algorithms designed for this purpose) may have to be communicated to staff.

An AI policy should be a living document that is updated whenever new systems are approved, or any significant changes are made. Regular checks can  help ensure the policy remains relevant and legally compliant. 

If this is an area you’re keen to explore further, feel free to reach out to me and the team here at CMS.