How to make AI serve marginalised people

Artificial intelligence (AI) has the potential for both helping and harming society. Like any technology, its impacts depend on the ways people choose to interact with it.

When it comes to digital inclusion, AI has proved relevant in a number of ways. It can positively or negatively impact the ability for people to participate safely online and to feel trust towards digital tools and processes. The concept of digital literacy has evolved to include AI, with the need to ensure people understand the changing impact of AI on their lives, and that they can use it in safe and beneficial ways.

In this article, we reflect on the AI Fringe 2023 and the events that have followed to explore some of the key ways to prevent AI from widening the digital divide.

Improve AI safety

According to the UK AI Safety Institute, introduced in January 2024, AI safety is “the understanding, prevention, and mitigation of harms from AI”. This relates to digital inclusion in that it addresses the ways that AI can contribute to a hostile, unsafe, or unsuitable online environment, particularly for those in marginalised positions or in situations of vulnerability.

Some known harms of AI include:

• AI being used for surveillance in ways that threaten the human rights of migrants and other marginalised people
• AI moderators in east Africa, India, the Philippines, and Lebanon experiencing psychological trauma from reading and viewing graphic content, low pay, and abrupt dismissals
• AI deepfakes being used to spread disinformation, undermine democracy, and harass and abuse women

Rachel Coldicutt, Executive Director at Promising Trouble and Careful Industries, told the Good Things Foundation’s panel at the AI Fringe that the most important action to take is holding policymakers and businesses to account in addressing AI’s harms. One way that charities can do this is through political activity and campaigning.

Unfortunately, according to the Charity Digital Skills Report 2023, 73% of charities say they don’t feel prepared to respond to the opportunities and challenges AI brings. Some key resources charity professionals can use to learn about AI ethics and safety are the Scottish AI Alliance’s Living with AI course and AI Safety Fundamentals’ Alignment Course.

Listening to service users’ experiences with AI harms can help charities represent their interests effectively to policymakers and businesses.

Ensure diverse meaningful involvement

A prominent theme in AI and digital inclusion is making sure that a diversity of people are involved in the development and use of the technology. The purpose of this is to reduce harm and maximise benefit for the most people possible, including those who are currently marginalised.

At the AI Fringe, Francine Bennet, Interim Director of the Ada Lovelace Institute raised the importance of centring the needs of people in the conversation when developing new AI products, as opposed to driving innovation through what is technically possible.

The event highlighted the unique power of charities and other civil society organisations to hold space for those at the sharp end of digital exclusion. Speakers identified a range of ways to do so, emphasising the need for collective action.

Helen Milner, Group Chief Executive at the Good Things Foundation, suggested that the sector could create its own, more representative datasets to address the biased data used by AI.

Multiple speakers urged the sector to channel stories of AI’s harms to policymakers and businesses, holding them accountable, and driving for justice and redress.

Bennet also highlighted the potential for a “more genuinely global” conversation about AI, where experts from the Global South are more central. This is necessary in part because of the unique opportunities and challenges that AI creates in countries in the Global South, which hold the majority of the world’s population.

Meanwhile, Lila Ibrahim, Chief Operating Officer at Google DeepMind, a Google team with the mission to “build AI responsibly to benefit humanity”, called for organisations to build diversity and inclusion into research, products, and programs from the very beginning.

Civil society organisations such as SwanseaMAD are focusing on AI literacy within communities with the express aim of diversifying the talent pool, lending to more inclusive participation in the design, development, and use of AI in the future.

Grow AI literacy

As AI’s use and influence grows, so does the importance of AI literacy throughout society. An element of digital inclusion, AI literacy has wide-reaching implications on personal safety and privacy, work and livelihoods, and civic participation.

One emerging framework of AI literacy suggests that most people should “know” AI, meaning they can copy, reproduce, recall, and memorise AI concepts, as well as “understand” AI, meaning that they can describe, explain, interpret, and demonstrate the meanings of AI.

The Scottish AI Alliance’s free Living with AI course answers the questions “what is AI?”, “where is it being used?”, and “how do I know I can trust it?”

SwanseaMAD’s AI courses for absolute beginners aim to teach AI skills in the way that sports are taught, by encouraging learners to have a go themselves.

Meanwhile, at Clear Community Web, teaching AI literacy involves demonstrating how it can help solve real-world challenges, such as forming a letter of complaint for a distressing housing situation.

Caspar Kennerdale, Managing Director of the CIC, reflects: “Any kind of change in technology does require us to bring people along with it, or more to the point, allow people to find their way through that journey themselves.”

“If we don’t give people the opportunity to participate, that divide’s always going to be there.”

Find out more

Our 2024 Digital Inclusion Summit revealed how charities can help can make the digital world a safer, happier, more inclusive place, from improving access to digital devices to demystifying cyber security. Click here to watch the session recordings for free.

Our report, ‘Digital inclusion in the UK charity sector’, uncovers charity practitioners’ attitudes towards digital inclusion, including the challenges charities face in reaching out to service users and how they are making the most of the digital technology they use. Click the link in the orange box below to download the report.

Our Digital Inclusion Hub features regular articles, podcasts, and webinars to help charities reach across the digital divide. Click here to learn more.