This is a fast-developing area of interest and concern for the VCSE sector. This resource aims to summarise the current picture and suggest further sources of information.

How is the sector currently using AI?

Currently available AI tools can do two main jobs;

• Collate and review large amounts of data faster and more accurately than people can.
• Use a combination of prompts and previous examples to generate ideas, text and images.

Current uses include;

• Chatbots (see glossary below) are providing 24/7 accessible legal advice, with oversight provided by trained professionals. 
• Predictive AI is being used to identify the people most at risk of becoming unhoused and put preventative measures in place. The same predictive powers can be used to identify increases in demand so that preparation can start in plenty of time.
• Many charities are currently using AI to generate social media content, for example putting together the text and images for a post on Instagram.

Things to think about before adopting AI

• AI makes mistakes. You must have skilled management and “prompts” (the instructions you give the AI) to make sure it does the job you want it to at a sufficient standard. You’ll also need to monitor constantly and fact-check to avoid issues for your organisation’s service and reputation. Competent ongoing AI use is never “cost neutral” and requires up front investment in things like upskilling staff.
• Using AI software such as ChatGPT, Google Gemini or Claude AI involves agreeing to sharing what you put on the site with other users and the world at large.
• There are some concerns that the general public doesn’t find AI a good fit with the level of personal service, and trustworthiness it expects from the VCSE sector.
• There are issues with data-protection with AI that can cause concern, particularly for clients and others who are sharing personal data with you.
• Many VCSE sector organisations have been resistant to adopting AI for ethical reasons. These include the very high environmental cost, the potential for AI to increase inequalities by stereotyping and potential data misuse. It’s important to reflect your values in the way you deliver your work, and to be open with clients and other stakeholders about the reasons for the decision you’ve made.
• AI adoption can look like the best solution to cost issues, but the fact that it is the current “hot topic” risks obscuring other ways to make things more efficient, including partnership and collaboration.
• Many AI tools are currently free, or embedded in subscriptions you may already be paying. However, as we’ve seen with other technologies that have become vital for everyday work life, once they are indispensable they often become charged-for. It is worthwhile bearing this in mind when assessing the long-term costs.

There are some general issues for the sector to also bear in mind

• AI adoption in other sectors is well ahead of the VCSE sector, and those organisations that have adopted it tend to be the larger, national charities with the time and resources to invest. If larger charities are able to show funders that they are using AI to save costs, they could become even more able to compete with the grassroots groups that may not be able to show the same efficiencies and will therefore need to ask funders to invest more.
• AI is often used to take over very basic, entry-level tasks due to the potential of saving in staff costs and time. However, this could lead to a skills-gap with fewer people possessing the basic skills needed to progress in a VCSE workplace AI may be used to replace roles currently held by volunteers, changing the  volunteering landscape.

In order to use AI well you’ll need the following;

• A clear answer to the question “why?”. It can be tempting to adopt an interesting new technology but it may well not be your best way forward.
• The skills in your team to select, commission, manage and use these new technologies effectively.
• Accurate costings and implementation plans.
• Access to up-to-date information about AI, to enable you to assess and revise your plans regularly in this fast-moving environment.
• Policy and procedures to ensure you are implementing AI in the safest and most impactful way to meet your charitable and social objectives.

Other sources of information

National Cyber Security Centre

NCVO guidance for small charities

Joseph Rowntree Foundation report on grassroots group’s perspective

Some jargon-busting info

Algorithm: A sequence of rules given to an AI machine to perform a task or solve a problem.

AI Ethics: the issues that AI stakeholders (your organisation plus engineers, government officials, etc.) must consider to ensure responsible development and use of AI.

Artificial Intelligence (AI): can be defined in many ways. It’s an umbrella term for a range of algorithm-based technologies that solve complex tasks by carrying out functions that previously required human thinking. Decisions made using AI are either fully automated, or with a ‘human in the loop’. As with any other form of decision-making, those impacted by an AI supported decision should be able to hold someone accountable for it.

Chatbot: A software application designed to imitate human conversation through text or voice commands.

Generative AI: a subset of artificial intelligence that uses machine learning techniques to generate content. It can create new content that is similar to the input data it has been trained on. Examples include creating images, writing text, and composing music.

Hallucinations: In the context of AI, they are the generation of information that isn’t present in the input data. This can occur when AI models, such as language models, generate outputs that include false or misleading information.

Large Language Model (LLM): a type of AI model designed to generate human-like text. It’s trained on a large amount of text data and can generate sentences by predicting the likelihood of a word given the previous words used in the text.

Machine Learning (ML): a subset of AI, providing systems the ability to automatically learn and improve from experience without being explicitly programmed.

Disclaimer

We make every effort to ensure that our information is correct at the time of publication. 

This is only intended as a brief summary of relevant issues and information. Legal advice should be sought where appropriate. The inclusion of other organisations in this information does not imply any endorsement of independent bodies, they are just for signposting purposes.

Voscur is unable to accept liability for any loss or damage or inconvenience arising as a consequence of the use of this information.