In February 2025, France hosted the third global artificial intelligence (AI) summit. Despite promises of strengthening an international dialogue to steer regulation of evolving technologies, it was ultimately a major lost opportunity for meaningful and honest discussions on AI’s risks and human rights harms.
For human rights defenders, the summit underscored that policymakers lack an understanding of and ignore how AI systems are harming people every day and how corporate greed and power are destroying our world.
Since the dawn of ChatGPT, which has led to an inflated AI hype phenomenon, policymakers are constantly influenced by narratives shaped by tech titans as a way of deflecting meaningful human rights-centric regulation.
For instance, in preceding AI summits in the United Kingdom and South Korea, existential doomsday narratives – primarily pushed by industry – dominated discussions. Equally, the Paris summit discussions, mainly driven by states, made AI sovereignty a central theme of the event, with officials promoting their countries as the “next place” to invest in AI while emphasising how an independent local technological landscape helps us achieve greater public good.
In both cases, the discourse was shaped by some powerful actors without addressing any human rights harms exacerbated by AI-driven technologies.
A charade of economic progress was also part of the mix during the French summit, promoted by the host country itself. During his opening speech, President Emmanuel Macron made his own sales pitch, focusing on France’s potential supremacy in AI development. Not once did the French president acknowledge the human rights harms posed by these technologies or the necessary steps toward binding regulation.
At a more global level, it was also alarming to witness J.D. Vance’s first foreign policy speech urge ever more AI deregulation. Similarly, the European Commission has recently made a U-turn from the promises it made about ensuring an AI regulatory framework based on human rights.
While the EU has made a name for itself as a champion of “trustworthy” and regulated AI development and deployment, the European Commission’s current agenda has seen a drastic pivot. Both Commission President Ursula von der Leyen and executive vice-president Henna Virkkunen promised the EU would cut red tape for business. The Commission has made good on its promise by axing the development of AI liability rules and downgrading corporate accountability rules.
Also read: AI Does Not Have the Answers to India’s ‘Aspirational’ and Frustrated Economy
The facts now clearly indicate that global competitiveness and independent sovereign tech capabilities are the key priorities for states, even if that means rolling back regulatory gains. It is unfortunate that this is happening despite countless investigations exposing how AI is amplifying biases and raising serious human rights concerns in areas such as policing, warfare, social protection, employment and more.
For instance, in the United Kingdom, Amnesty International has exposed how predictive policing systems are supercharging racism by unfairly profiling immigrant communities under the guise of “crime prevention.”
In New York City, a police-owned facial recognition system is eroding civil rights to privacy and disproportionately impacting marginalised groups, including Black Lives Matters protestors.
In warfare, autonomous weapons systems operate without any human oversight, functioning as mass-killing machines.
In welfare provision, algorithmic decision-making is depriving people of necessities like food, healthcare and shelter.
The list of harms is endless – just like the number of people affected.
The next summit will be hosted in India, and this is also a cause for concern. Given the country’s track record of abusing AI technologies, it’s unlikely that human rights regulation will be high on the agenda.
Amnesty International has documented how the use of a facial recognition system in the city of Hyderabad is leading to mass surveillance, making it impossible to walk down the street without risking exposure to facial recognition technology.
Following media reports, Amnesty International also exposed how an automated system called Samagra Vedika in the Indian state of Telangana is preventing eligible people from receiving welfare.
During the next summit, state officials must acknowledge that we are now at a critical juncture where we need to counter the uninhibited push by both billion-dollar corporations and state officials to fuel unchecked AI economies and build extensive AI-driven state machinery by embracing deregulation and monstrous AI investments.
Governments must recognise that the need for a human rights-centric roll-out of these technologies is more urgent than ever and that fundamentally, this means that if a tech system is perpetuating human rights harms, it should never be deployed.
In a rapidly changing global environment where we are witnessing a backlash against civil liberties every day, states must confront and tackle the harms caused by AI instead of writing the rules of our future with tech giants.
They must respond to calls from civil society groups and impacted communities and work together to ensure that corporations are subject to enforceable and binding AI regulations that protect us all.
The author is senior director of research and policy, Amnesty International.
