At the third global artificial intelligence (AI) summit in Paris, governments and companies made bold promises about creating AI that serves the public good. Yet, experts are raising alarms over the underlying tensions. Many attendees expressed concerns that while there’s a push for AI to be an open and public asset, it’s still largely controlled by a small number of powerful corporations and countries.
Despite the rhetoric around inclusivity, the pressure for deregulation is palpable. Some key figures in politics and industry seem more focused on easing rules than on ensuring safety and building public trust. They argue that regulations slow down innovation, but this raises questions about who truly benefits. For AI to genuinely serve the public interest, the lack of proper guardrails could lead to diminishing safety and ethical standards.
The summit came on the heels of the UK’s AI Safety Summit and the second AI Seoul Summit, both of which highlighted the safety and risks associated with AI. In contrast, the AI Action Summit covered broader topics like public service applications, the future of work, and global governance, bringing attention to pressing issues surrounding AI development.
Over the two days, the summit announced two significant initiatives: the Coalition for Environmentally Sustainable AI and Current AI—a foundation backed by multiple governments and companies, including Google and Salesforce. Current AI aims to drive technology toward social benefits. This initiative also seeks to expand access to quality datasets, enhance transparency, and measure AI’s social impact.
European governments pledged around €200 billion in AI-related investments, marking the largest public-private investment globally. Alongside this financial commitment, representatives from 61 countries signed a statement promoting inclusive and sustainable AI, which aimed to bridge digital divides and reinforce international cooperation.
However, the UK and US opted out of signing the joint declaration, with UK officials stating they would only support measures aligned with national interests. Throughout the summit, leaders from the US and Europe echoed concerns about regulatory restrictions stifling innovation, calling for a simplification of rules to foster AI development.
After the summit, the EU decided to scale back its AI liability directive, sparking fears about a potential rollback on necessary safeguards meant to protect individuals from harmful AI decisions. The UK’s AI Safety Institute also rebranded itself to focus on security, sidelining issues of bias and free expression.
Many experts voiced concerns over the potential direction for AI development. They pointed out a disconnect: politicians want open and sustainable AI, but at the same time, they advocate for fewer regulations while investing heavily in technology without clear oversight.
Sandra Wachter, a professor focusing on AI ethics at the Oxford Internet Institute, noted the complexity of the regulatory debate. She questioned the existence of any laws truly stifling progress, arguing that most AI issues stem from the technology itself and the way it’s designed and deployed. Instead, she challenged the narrative that regulation hampers innovation.
Linda Griffin, a Mozilla executive, also cast doubt on anti-regulation arguments, emphasizing that the drive for profits in a few big tech companies doesn’t equate to benefits for everyone. Gaia Marcus from the Ada Lovelace Institute highlighted the need for governments to incentivize safer, more trustworthy AI systems rather than let corporate interests dominate the conversation.
Despite the summit’s broader discussions on public interest AI, clarifying the definition of “public interest” remains crucial to avoid capture by narrow corporate agendas. Nyalleng Moorosi from the Distributed AI Research Institute stressed the importance of transparency and inclusion in building AI systems that truly reflect community needs.
Amid hopes for more collaborative efforts, many experts agreed that the intense market concentration in AI is a significant issue. Wachter pointed out that the political discourse around AI is largely shaped by the interests of a few dominant players, leaving many voices unheard.
Looking ahead, it’s clear that while discussions to reform the AI landscape are underway, the ultimate challenge remains creating a system that genuinely benefits a broad spectrum of society rather than a select few. Conversations around open source AI are gaining traction as a means to address concentration concerns, though it remains to be seen how effectively these initiatives will reshape the market dynamics in favor of public benefit.
