How Should DEI Leaders be utilising Generative AI?

Image description: A photograph of Hayley at work. Hayley is wearing a white blazer and has her hair in braids. She is looking down at her laptop typing. The HTVB logo is in the background.

The ethical concerns associated with AI been widely discussed, as well as the lack of diversity within teams working on AI and the corresponding data sets used to train tools. However, I am yet to see DEI leaders discuss the implications of using AI in our work - which we know is happening - even if people are not admitting it!

Be honest if you are working in DEI and are utilising AI in your day-to-day work. AI, specifically generative AI, is now part of our world in ways we probably did not imagine it would be. In this blog post, I am sharing some thoughts, based on what I have noticed, on what I see as the Do's and Dont's for DEI leaders when it comes to using generative AI to support our work. 

Generative AI, has the ability to transform the way we communicate and potentially even enhance the inclusivity of our communication. By leveraging AI, DEI leaders can ensure that their messages are tailored to their audience and are delivered in a way that is easily understood. Just like any tool, it must be used thoughtfully, taking into account its constraints and possible drawbacks. 

Do:

1. Ask AI tools to help you ensure your written communication is clear and concise.
2. Use AI apps to help you on your administrative talks that help you save time for your more meaningful work.
3. Use AI input to make a communications skeleton plan to help your organisation to commemorate key dates.
4. Try AI functions on presentations to help  create visual aids that support education and engagement.
5. Aid information retention by enabling AI tools in Teams, Zoom etc, helping make your meetings more effective and inclusive through clearer notes.

Don't:

1. Rely only on AI for creating translations and transcriptions for your colleagues. Why? Whilst translation and transcription technology has come a long way, it is still riddled with mistakes and inaccuracies that make understanding content a barrier for people you are trying to support! Read about the dangers AI translation can pose in the case of asylum applications.
2. Misuse the trust of colleagues by inputting their data into AI tools. Even if it's anonymised, this is unacceptable! Why? It is not ethical to input colleagues' data into AI tools, especially when you have not told them this will be done. Misleading people in this way can breach trust within the team and violate individual privacy. 
3. Use email writers that can stop you from meaningfully connecting with people, and make you come across inauthentic. Why? Relying too heavily on AI-generated content will undoubtedly make you seem inauthentic. As a DEI leader, you need to role model authenticity and openness which may contradict with using AI in your writing - for some people. It is all about balance, so if you are using AI to speed up your writing, think about how you could come across, and whether it is appropriate in that context/to that audience.  
4. Rely on AI generated content to be effective in raising awareness. Why? While AI can assist in creating content that is engaging, and could help raise awareness of DEI topics, it should not replace the human element. Evidence tells us that inter-group contact (aka contact between members of diverse social groups) is the effective way of creating awareness and educational outcomes we strive for in DEI work. This cannot be replaced by AI. 
5. Create a DEI strategy or workshop using Chat GPT recommendations. Please just don’t! Why? There are so many reasons for why this is not okay (see my last post for what makes an effective DEI strategy!) but most of all, we have to remember that AI still has a huge potential for error. The responses you receive are not likely to be based on evidence, and could cause more harm than good! 
6. Use AI generated images of ~ diverse ~ teams to give a misleading impression of your workplace. Why? Other than the reality that AI image generators often give racist and stereotyped results, these creepy images overstate how diverse your team is which gives an unfair and false impression, causing harm to marginalised folks internally and externally. 

Author - Hayley Bennett, Director of HTVB.