Part 5: Conclusion and Recommendations

Despite facing threats of surveillance, censorship, disinformation, and a civic space under attack around the world, CSOs have been fighting against harmful trends and patterns in AI adoption by governments and the private sector. On the global stage, they have been successful in advocating for a ban on fully autonomous weapons at the UN Convention on Certain Conventional Weapons¹⁵¹ and campaigns led by tech workers have educated the public on the complicity of Big Tech in Israel’s AI-assisted genocide in Gaza and occupation of Palestinian Territory.¹⁵²

In Brazil, civil society efforts have been successful in advocating for a more robust and human rights-based alternative law to a weaker and self-regulatory bill that was favoured by the private sector. Across the world, civil society organisations have been building their own small language models and their own datasets to advance AI uses that align with their values and respect of human rights, including the right to privacy, free expression, and freedom from discrimination.

Despite the limited resources, civic space actors demonstrate that an alternative approach exists in contrast to the extractive nature of the AI models proposed and implemented by Big Tech and then used by governments. However, for civic space actors to leverage AI’s inclusive, non-extractive, and beneficial aspects for all, and to prevent its harmful applications, the sector needs resources and capacity building.

• . With financial resources, CSOs can hire technical expertise to help integrate AI in their work, build their own models and datasets, and invest in the needed infrastructure, both software and hardware.
• Networks and initiatives that support CSOs can also strengthen the capacity of CSOs, including starting withbasic knowledge and understanding such as demystifying AI, the benefits and drawbacks – to deploy it ethically and in alignment with human rights.
• There is a need for multilingual toolkits, resources, and templates for frameworks and policies that can help guide CSOs in their deployment of AI. These resources can cover, for example, policy templates on AI use so that CSOs are transparent with different stakeholders and detailed technical guides on conducting human rights impact assessments when deploying and developing AI. Such resources should be available in as many languages as possible to reach as many organisations as possible around the world.
• In addition to guides and templates, many CSOs lacking technical expertise and resources will need mentorship and guidance from more technical coalitions and organisations to integrate AI in their work responsibly and assess and address human rights risks.
• It is essential to invest in tools, resources, capacity building and legal protection to allow CSOs to resist and challenge the dominance of Global North commercial models that do not reflect the realities in which they work. For instance, investing in independent initiatives that support AI sandboxes, the building of open, diverse and ethical datasets, and small language models could help CSOs access and deploy AI in ways that do align with their values and do not cause harm.
• In the face of AI-powered surveillance, censorship, and disinformation eroding civic space, it is worth exploring building and strengthening broad and diverse coalitions to exercise public pressure on governments and tech companies.
• It is also essential to continue raising awareness and building digital security skills of civil society actors to reduce risk of unauthorised access to data and mitigate such their impacts, particularly with emerging threats associated with AI.
• Given the far-reaching social, economic, and political implications of AI, the coalitions of CSOs pushing for robust AI tech global governance need to include more voices from the Global South and v and from other sectors such as unions, artists, and environmental and anti-racist movements.