Prohibition of Lethal Autonomous Weapons Systems Bill 2026: Second Stage
Seanad Éireann — 2026-06-17
============================================================

Patricia Stephenson (SD), Labour Panel

I thank the Minister of State for being here. I welcome the Bill being brought forward by Senator Alice-Mary Higgins and my Civil Engagement Group colleagues. It is powerful and pertinent legislation. It is calling for a moratorium on the design, production, sale and use of autonomous weapons. It is in line with calls from the UN to introduce international legislation to ban their use outright.
These weapon systems, which include advanced drones to target and apply force, operate without any human oversight. They are programmed to hunt and they are programmed to kill. They lack empathy and the ability to make ethical judgements, and, ultimately, they are unaccountable under international law. As we have heard before, the Secretary General of the UN has called these weapon systems morally repugnant and said they have no place in our world.
The threat these weapons pose to human life and human rights cannot be overstated. Increasingly, we are seeing the reduction of meaningful human control in the design and implementation of AI systems as the technology continues to evolve. This technology only serves to accelerate our own digital dehumanisation and we must, therefore, work collaboratively on an international level to halt the race to the bottom in the evolution of autonomous technologies. This is what we are seeing now: a race to the bottom to have the most effective AI technologies which are used to murder civilians.
I would not normally cite the Pope, but as we have already heard, he has recently called for the disarmament of AI, citing his serious concerns that AI poses an extreme danger to humanity and lowers the threshold to war. This is a warning that should be alarming to all of us and we must heed it. Since 2014, UN member states have been debating proposed regulations on autonomous weapon systems under the Convention on Certain Conventional Weapons, but so far there has not been a consensus on the definition of "lethal autonomous weapon systems". There has not been a clear roadmap on how to effectively legislate for what constitutes meaningful human control. We must, therefore, act domestically. International agreements are the best solution, but in lieu of that, we must have our own domestic legislation. As was said earlier, Ireland could lead on this issue. We could be the first to do it and then have it rolled out elsewhere.
Since 2014, in the meantime, the IDF has used AI technology to target and kill Palestinian civilians. The Lavender AI programme, very sweetly named, automatically scans surveillance databases to generate the names of thousands of suspects for targeting. Lavender can identify targets in just 20 seconds, without any human oversight at all. The IDF uses the programme Where's Daddy? It uses AI technology to track and target Palestinians through their mobile phones. It then marks the exact moment when they return home to their families. Daddy comes into the house and the Where's Daddy? technology triggers an attack, usually at night. Entire families have been targeted and murdered using this form of weaponised AI.
Far from being accidental casualties and a consequence of AI's tendency to get things wrong, these practices are designed to deliberately kill as many people as possible with minimal personnel needed. Rather than minimising harm, these AI tools have effectively turbocharged the pace and scale of deadly force but repackaged it under the guise of algorithmic efficiency, often against civilians. What was once decision support by AI has evolved into decision automation and the effects on human lives are devastating. Across the board, we as legislators are constantly sprinting to keep up with the rapid evolution of AI technology.
Where appropriate, AI technology should be harnessed to support human decision-making that promotes well-being and positivity in society. However, in the use of AI in autonomous weapon systems, the human element is entirely removed. People cannot reason with an algorithm or an AI weapon. They cannot question it or talk down to it, and drones cannot be tried in a court of law. What is worrying is the opacity with which artificial intelligence is learning and making decisions. In many cases, we simply do not know how AI is making automated decisions. It operates within this black box. We have no oversight of it at all. AI is really good at crunching numbers and analysing vast volumes of data quickly but it struggles in situations where context is essential. Another worry with AI is the inherent discrimination built into its systems, which we cannot discern. The question is: who is programming the algorithms?
AI is superior in one way, in that it does not suffer from cognitive bias but when it is fed biased data, which is always the case, it takes that bias and runs with it. Concerns around bias have come up in the policy debates on autonomous weapons but the topic has not been considered or explored in depth. It is important to understand how bias in autonomous weapon systems threatens compliance with international humanitarian law, and the legal implications of this have to be addressed. We have heard chilling reports of drone strikes on civilians in Ukraine, Gaza, Lebanon and Iran. What is frightening is states' inability or unwillingness to draw a red line that states that using drones in AI to target humans is wrong and a step too far.
The Campaign to Stop Killer Robots notes that the vast majority of states are in favour of introducing binding legislation to regulate and prohibit autonomy in weapons systems. As far back as 2019, Irish people were cognisant of these harms, when 81% of people surveyed were in favour of a ban.
There is still a lack of understanding of how these weapon systems work and it is crucial to emphasise the existential threat posed by the proliferation of the autonomous weapons. We need to devise solutions to restrict their use. We have grown accustomed to AI recommending podcasts we might like to listen to and certain products it pushes us to buy, and we have serious concerns about that, but AI is also recommending human targets to the US and Israeli militaries, and militaries all over the world. It is recommending different people that they can kill. Anthropic's large language model, Claude, assisted the US military in recommending strikes in Iran. AI recommended the targets and the soldiers pulled the trigger. More than 1,000 targets were struck within a 24-hour period, including many residential areas. That is violent on an unprecedented scale.
The US military uses AI technology to save time in identifying as many targets as possible. Human beings are simply reduced to data points. They are sorted, scored and processed for elimination. Elimination, of course, is death. It sounds dystopian but it is happening today. It seems that Iran is a laboratory and the US and Israel are acting as scientists, testing their toys on a civilian population. We also know there is significant money to be made in this industry. It is a big money-making industry and introducing legislation to ban further development and use of autonomous weapons will probably not go down with major tech and military companies but any rules to safeguard human lives have to be prioritised.
The risk to human life and human rights is too great and it is for that reason that I hope the Government will not only not oppose this Bill, but seek its quick processing through both the Seanad and the Dáil. I note a point made by Senator Higgins in 2024 at the Joint Committee on Education, Further and Higher Education, Research, Innovation and Science when she was scrutinising legislation on Horizon funding, which is that we are seeing Horizon funding support cross-fertilisation into these types of civil and defence dual-use technologies and AI technologies. We therefore need to explore where that funding is going and whether it is going in to support the use of automated weapons.

---

Source: Houses of the Oireachtas. Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). The Official Report is revised after first publication; the fetch timestamp below identifies the version quoted.

Record URI: https://data.oireachtas.ie/akn/ie/debateRecord/seanad/2026-06-17/debate/main
Retrieved: 2026-08-14T04:55:04+00:00
Sitting date: 2026-06-17