In 2023, I wrote The Human Rights Risks of Artificial Intelligence (AI). At the time, some of the concerns around AI still felt theoretical. We warned that AI could reinforce discrimination, erode privacy, harm workers and creators, concentrate power, spread misinformation, and contribute to environmental harm.
Three short years later, many of those risks have become reality. In fact, in some cases, the reality is more concerning than we anticipated.
Here, we’ll revisit our concerns about AI and look at how they have played out since.
AI in Warfare
The risk: In our original article, we naively stated that the human rights risks of AI were “often more subtle than the headlines about autonomous robots and weapons systems might suggest.”
The reality: Unfortunately, those more obvious risks have become reality too. AI systems are now being used to help select targets in warfare, including in Gaza. What once sounded like a frightening future scenario is happening now.
War already results in devastating violations of human rights—loss of life, the displacement of communities, the destruction of homes and essential infrastructure, and lasting trauma. Staff at our parent organization, The Center for Victims of Torture, know this devastation well.
Ultimately, AI has not reduced warfare or its human cost. It has increased our capacity to wage it faster and at greater scale.
For human rights defenders, the impact is devastating: technology can now accelerate the very violence, displacement, destruction, and loss of rights that our work seeks to prevent.
Privacy and Surveillance
The risk: We warned that AI could erode privacy by collecting and analyzing enormous amounts of our personal movements and information. We also raised concerns about AI-powered surveillance being used for repression and authoritarian purposes.
The reality: We are seeing how quickly surveillance technology can be widely implemented—and how easily it can become subject to abuse and error.
Flock license plate readers, for example, allow law enforcement agencies to automatically record and search people’s vehicle movements, creating vast networks of location data. That surveillance power has already been abused. The Institute for Justice has documented multiple incidents of law enforcement officers using automated license plate reader systems to stalk current and former romantic partners.
And even when these systems are used as authorized, they can be alarmingly inaccurate. Investigation found significant errors in Flock alerts, including in one California town where 71% of the alerts examined were inaccurate.
The human rights concern goes beyond whether these systems are accurate or whether individual officers abuse them. Privacy itself is a human right. Article 12 of the Universal Declaration of Human Rights protects people from arbitrary interference with their privacy, family, home, and correspondence. Widespread systems that continuously collect and make our movements searchable raise serious questions about whether the surveillance itself violates that right.
When those same systems are then misused to track individuals—and produce inaccurate results—the human rights impact becomes even more alarming.
Workers and Creators
The risk: In 2023, we warned that AI could threaten workers’ rights and livelihoods. We were particularly concerned about artists whose work was being used to train AI systems without consent or compensation.
The reality: AI is now competing directly with the people whose work helped make these systems possible.
AI training facilities are buying books, scanning them for data, and destroying some of the physical copies. Recent reporting has raised concerns about companies purchasing rare and antique books for this purpose.
The same tension is playing out in music. Some AI-generated songs have reached the top of music charts and gone viral on social media. As catchy as the AI-generated “First Time in San Juan” trend was, we have to ask: What happens to human creativity when machines can reproduce the products of human creativity without sharing the costs of becoming a human artist?
This is more than a copyright issue. It is a human rights issue.
The Universal Declaration of Human Rights protects our right to decent work and a livelihood, but it also recognizes our right to participate in cultural life and protects the interests of people who create artistic work. Human beings have always made music, art, stories, and other creative works—not only to earn a living, but to experience joy, express ourselves, preserve culture, and connect with one another.
If we automate the creative work people love while leaving humans to simply consume, produce, and work more efficiently, what exactly are we making all of this efficiency for?
Human progress should create more space for people to live, create, rest, and connect—not reduce us to more productive cogs in the machine.
The Right to Truth
The risk: In our original article, we focused heavily on the right to truth. We warned that deepfakes and synthetic media would make it increasingly difficult to distinguish reality from propaganda and falsehood.
The reality: That problem has grown—and it has taken a turn I did not fully anticipate.
AI can create convincing fake images, videos, audio, and text, making it increasingly difficult to separate fact from fiction. But there is another side to this problem that I did not fully anticipate in 2023: AI doesn’t just make it easier to pass off something fake as real. It also makes it easier to dismiss something real as fake.
“If something happens really bad, just blame AI.” – U.S. President Donald Trump
We are seeing a concerning trend in which political and global leaders use the existence of AI-generated content to cast doubt on authentic evidence, including documentation of violence and human rights violations. Researchers call this the “liar’s dividend”: once we know convincing fakes are possible, those in power gain another way to deny or discredit real information. For human rights defenders who rely on photos, videos, testimony, and other evidence to document abuses, the consequences are serious. If authentic evidence can simply be dismissed as AI-generated, how do victims prove what happened? How do we hold perpetrators accountable?
The Universal Declaration of Human Rights protects our right to seek and receive information (Article 19) and our right to an effective remedy when our fundamental rights have been violated (Article 8). Both depend, in part, on our ability to establish what happened. AI-generated misinformation—and the growing ability to dismiss real evidence as AI—can make documenting human rights violations, seeking justice, and holding those responsible to account even harder.
Power and Accountability
The risk: In 2023, I asked whether technology companies really needed more power and profit. We warned that AI could concentrate wealth and decision-making power in the hands of a few technology companies.
The reality: That concentration has only become more visible. Building the largest AI systems requires enormous amounts of computing power, data, energy, and money. As a result, a relatively small number of corporations have tremendous influence over technology that increasingly shapes how we work, communicate, create, and access information.
At the same time, AI systems are increasingly being designed to carry out complex, multi-step tasks with less human intervention along the way. In 2026, for example, AI models being tested on cybersecurity tasks accessed real companies’ systems after a supposedly isolated testing environment was mistakenly connected to the public internet. This incident shows how quickly AI systems can cause unintended consequences when human safeguards fail.
Accountability ultimately rests with the people and institutions that create and implement these systems. We have already seen that we cannot rely on AI companies to voluntarily put human rights ahead of growth and profit. This makes meaningful government oversight and regulation essential to protecting the people and communities affected by these technologies.
Yet we are already seeing that responsibility challenged. Communities across the United States are pushing back against AI data centers over concerns about water and energy use, pollution, rising utility costs, and other local impacts. President Trump warned that communities rejecting data centers could become “backwards and poor” and urged Americans to “let Data Reign.”
Who is protecting who here? Governments have a responsibility to protect human rights, including from harms caused by private companies. Yet we are seeing a troubling trend of governments prioritizing rapid AI expansion even as communities raise concerns about the human and environmental costs. This growing alignment between government and an already powerful AI industry demonstrates exactly why we need strong, enforceable guardrails. As corporate power grows, so must public accountability and human rights protections.
The Environmental Cost of AI
The risk: We warned that AI’s environmental impact could disproportionately affect vulnerable communities. Data centers require enormous amounts of energy and water, even though we often think of AI as something that exists somewhere “in the cloud.”
The reality: There is no cloud. There are buildings, servers, power plants, cooling systems, water supplies, and real communities supporting the infrastructure behind AI. As demand for AI grows, so does this physical footprint. Data centers can consume enormous amounts of electricity and water, while nearby communities may face increased utility costs, competition for local resources, pollution, and other environmental impacts.
Once again, we need to ask: Who benefits, and who bears the cost? The United Nations recognizes a clean, healthy, and sustainable environment as a human right, while the Universal Declaration of Human Rights protects the right to a standard of living adequate for our health and well-being. When the environmental costs of AI fall disproportionately on marginalized communities, while its wealth and power are concentrated elsewhere, this is not simply an environmental concern. It is a human rights concern.
Human Rights Must Come First
When I wrote the original article in 2023, I ended with a question about the kind of world we want to live in. I still think that is the right question:
Do we want to live in a world where our everyday interactions with other human beings become automated by systems like AI? I certainly would prefer going in the other direction, a direction that values our connection to one another and the planet.
The risks we identified in 2023 were not inevitable, but they were real. In just a few years, we have seen AI contribute to surveillance, misinformation, military targeting, exploitation of creative work, concentrated corporate power, and environmental harm. Many of these technologies have developed and spread faster than the protections needed to prevent their human rights impact.
We need to stop treating technological advancement as automatically equal to human progress. Progress should be measured by its positive impact on people—whether it protects our rights, strengthens our communities, expands our freedoms, and improves our lives. As I wrote in 2023,
“There are already enough systems working against the world’s most vulnerable. We must be sure with every new system put in place, that we are protecting the human rights and dignity of all who are affected.”
Technological solutions to the world’s problems must abide by the principle of Do No Harm. They should not make inequality, exploitation, surveillance, and violence faster or more efficient.
Human rights give us a framework for deciding what kind of future we are willing to accept. The rights to privacy, information, decent work, cultural participation, health, and human dignity are not negotiable. They belong to us simply because we are human—and that does not change because a new technology is powerful or profitable. Human rights must be protected throughout every stage of technological advancement.
So, are we doomed?
We don’t think so. Around the world, activists, organizers, artists, researchers, lawmakers, and everyday people are already questioning these systems, exposing their harms, demanding accountability, and finding creative ways to resist. The same human ingenuity that created these technologies can also be used to challenge them.
The future of AI is not inevitable. We still have a say in what gets built, what gets regulated, what becomes normalized, and what we refuse to accept.
In Part IV of this series, we’ll look at some of the tactics giving us hope—and what all of us can do to help build a future in which technological advancement serves people and protects human rights.
Read the earlier articles in this series:
