How Paris AI Summit was a quiet success – Firstpost

It is simplistic to dismiss the outcomes of the Paris AI Action Summit as underwhelming or disappointing. Yes, the US and the UK declined to sign the final statement, underscoring the challenges in achieving a global consensus on AI governance. But this was not unexpected and aligns with Trump’s ‘America-first’ policy. US Vice President JD Vance said as much, “The United States of America is the leader in AI, and our administration plans to keep it that way.”

It is concerning that the only agreement between the US and China, the world’s foremost leaders in AI today, is that the decision to use nuclear weapons should be under human control, not AI.

Unlike the previous two summits in Seoul and the UK, the Paris Summit’s final document is a humble ‘Statement’, not a grand ‘Declaration’. The reduced scope and scale of the ambition should not distract us from the fine print that the summit was able to negotiate and advocate for quietly.

The Paris Summit shifted away from the previous summits’ focus on AI safety and risks. While necessary to mitigate AI’s potentially catastrophic implications, the sole emphasis on safety’s role in avoiding harm was abstract and allowed free rein for doomsday thinking that was counterproductive to the credibility of the Seoul and especially the Bletchley Summit. The spotlight on the long-term negative consequences of AI also shifted the focus from the leverage that policymakers have today to shape the technology’s incremental, here-and-now implications.

The Paris Summit brought back global attention to the real implications of AI for the people and economies affected by it. Titled the AI ‘Action’ Summit unlike the previous iterations, the Paris Summit discussed AI’s impact on the future of work and the labour market. It highlighted the need for increased consumer protection and intellectual property. It railed against the concentration of the technology in a few firms or countries and promoted open-source AI for increased trust and transparency.

The statement is vague in its wording – it calls for “inclusive” AI and to “reinforce the diversity of the AI system” – but the message is clear. It advocates for a bottom-up development of AI that is more wide-based and democratic. This is an essential counter to the increasing politicisation of open source, with the US considering a ban on the export of model weights.

The summit is noteworthy for its efforts in AI multilateralism, especially amidst increased geopolitical tensions. The statement mentioned the UN General Assembly’s Global Digital Compact and advocated for an “inclusive, multi-stakeholder dialogue and cooperation”.

Touted the ’third way’, the EU and France committed $320 billion to rival the US and China – a definitive step which is a clear departure from the EU AI act that prided itself in being the “first comprehensive regulation on AI by a major regulator anywhere”. This is a signal that the EU is scaling back red tape believed to have hindered its technological progress and finally pushing instead for innovation, growth and progress.

Significantly, the Paris Statement views technology not as an end in itself but a means to a better, more prosperous life for people globally. It stressed the need and urgency to “assist developing countries in AI capacity building” and called for the setting up of a “Public Interest AI Platform and Incubator”, a global platform that would support AI projects necessary for the public good, championing open-access AI while maintaining national AI sovereignty.

Fitting for a city known for a climate agreement, the summit sought to promote ’energy-friendly’ and ‘sustainable’ AI in light of AI infrastructure’s increased energy demand and its estimated fossil fuel burn. While ‘climate change’ was left out in the final statement, France used the opportunity to promote its nuclear capabilities and push for nuclear and electricity-powered AI.

Even though the statement is non-binding and lacks enforceable regulations, it is nevertheless a positive direction for the global AI landscape.

The Paris Summit had the largest representation from nation-states of the three AI summits, with 60 signatories of the final statement compared to the Bletchley Declaration’s 29.

It should be remembered for bringing clarity to the global AI debate, contextualising it for the real pains and aspirations of people today, and seeking to gather coordinated support to counterbalance the heft of the US and China’s self-serving attitudes.

If one were to measure by the conspicuous absence of the US’ endorsement, the summit would not be a resounding and chest-thumping success. However, in its subversion of previously set agenda and narratives around the broader AI governance modalities and in its size and scale that left the US and the UK to watch from the sidelines, in its stand for an open, democratic AI that is tailored to technology sovereignty needs of nations today, it would have left its mark.

Shobhankita Reddy is a research analyst at the High-Tech Geopolitics Programme of the Takshashila Institution. The views expressed are personal and not that of Firstpost.