How to protect the most vulnerable in the age of AI

Pope Leo XIV’s warning that artificial intelligence must be developed with accountability, delivered via a landmark encyclical, came at a conspicuous time. As AI capabilities become more accessible, criminal actors are incorporating them into existing forms of fraud, trafficking and online exploitation. These developments underscore the importance of ensuring that innovation is accompanied by effective safeguards and accountability mechanisms.

The pope's warning reflects a broader concern shared by policymakers, business leaders and civil society organizations: that governance and accountability must evolve alongside technological capability.

Amid the AI euphoria, the challenge is to ensure that AI's enormous potential for economic and social benefit is matched by equally robust efforts to prevent misuse and protect vulnerable communities.

Every AI tool that can explode the growth and productivity of good and legitimate enterprises can do the very same for the most brutal and abusive enterprises, and AI tools are already being used to do so.

The warning signs are vivid for me. As Founder & CEO of International Justice Mission (IJM), I’ve spent the last 30 years working with teams partnering with justice systems around the world to protect the most vulnerable citizens from violence and abuse. Our work keeps us uncomfortably close to horrific criminal behaviour and gives us a unique window into the way AI is already empowering the most brutal criminal endeavours.

From our teams in Southeast Asia, we know that criminal gangs used AI-empowered technologies last year to steal $50-$80 billion from unsuspecting citizens around the world through industrial-scale online scamming operations powered by modern slavery. Thousands of trafficking victims are confined in guarded compounds where they huddle over workstations in 24/7 shifts running continuous online scams. IJM has supported thousands of survivors from these compounds, and they tell us how they faced beatings, starvation, electric shock and sexual assault if they didn’t meet their scamming quotas.

Last year, these scamming operations generated profits that were significantly larger than the combined national government budgets of Myanmar, Cambodia and Laos, where these operations largely reside. AI-empowered technologies supercharged these criminal enterprises through capabilities like AI translation tools, voice cloning, deepfake videos and hyper-scalable recruiting technologies.

Another brutal and exploding tech-enabled criminal enterprise is livestreamed child sexual exploitation, and once more, this is made worse by AI. Sex offenders in places like the US and UK pay a ‘facilitator’ to direct the live sexual abuse of a child in front of a webcam on the other side of the world. IJM research recently exposed that half a million children in the Philippines were sexually abused on online platforms in a single year.

AI is also transforming and accelerating traffickers’ recruitment of victims into forced labor industries across the world. Fraudulent job advertisements powered by AI can now be translated into even the most obscure languages, targeting people who would never have been reached before. AI-powered chatbots create convincing fake job postings at scale, tailoring deceptive messaging to specific demographics.

The good news is, we know what to do. Pope Leo’s call in his encyclical for “robust legal frameworks, independent oversight, informed users and a political system that does not abdicate its responsibility” is the kickstart we need as global citizens. In every era, new technologies have emerged that could be used to accelerate productivity and human flourishing and simultaneously misused to devastate lives. Pharmaceuticals, nuclear energy and high-speed transportation are obvious examples. In each era, leaders did two things to successfully safeguard the future: they took responsibility for the dangers being created and built accountability for those who chose not to.

Business leaders must take responsibility. As tech leaders create AI tools and grow their capacity, they must take responsibility for understanding how criminals will use them, consult with law enforcement and be as clever and passionate about mitigating their abuse as they are at maximizing their use.

They should also commit to “safe-by-design” standards for the devices they manufacture, to stop exploitative use before it happens. For instance, on-device detection technology (powered by AI) can detect when a device is trying to record a pre-pubescent child exploited in a sexual act and simply not allow this universally illegal content to ever be created.

Governments must establish accountability. No other industry has ever been allowed to roll out whatever powerful product it wants to the public without eventually facing regulatory guardrails against serious public harm from misuse, abuse or criminal use. The enduring regrets of history have been innocent lives destroyed and wasted while those guardrails were neglected. There is no sane reason why the AI technology sector should be granted a singular exemption from this commonsense lesson of history. Indeed, without regulatory requirements, corporate leaders privately acknowledge that market dynamics will drive toward removing guardrails, even those protecting children.

The truth is, we know more than enough to take responsibility now.

As the UN’s chief investigator for the genocide in Rwanda, I know what it feels like to show up too late to stop a catastrophe.

The experience taught me the importance of acting on early warning signs. AI presents extraordinary opportunities for human progress, but it is also creating new avenues for exploitation. The challenge before us is to ensure that governance, accountability and protection keep pace with innovation.

The lesson of Rwanda is not one of inevitability, but of prevention. Societies are strongest when they recognize emerging risks early and respond with urgency, cooperation and foresight. The rapid evolution of AI gives us an opportunity to do exactly that.

I believe history will one day convene a tribunal of our grandchildren, and they will simply ask us: where were you when the critical choices for the future of AI and humanity were being made? I hope we can say that we showed up to take responsibility and to build accountability to secure their future. And that we showed up on time.