Should a police drone be allowed to carry a weapon? That question is no longer hypothetical. In the United States, companies are already selling drones to police departments, and some are signalling that they will not restrict how their technology is used. A growing number of voices are calling for lawmakers to act before armed drones become routine.
For Irish business owners, this matters because the same questions are being asked in Europe. The EU AI Act, which came into force in stages through 2025 and 2026, classifies certain AI systems as high-risk. Autonomous drones that could cause physical harm fall squarely into that category. Understanding how Europe is approaching this issue helps you anticipate what regulations may affect your own use of AI.
The situation in the United States
The Electronic Frontier Foundation (EFF) has been tracking the issue of armed police drones since 2021. Their recent report highlights two worrying developments. First, Skydio, one of the largest drone vendors to US police departments, has signalled that it will not restrict customers from arming their drones. Second, a company called Campus Guardian Angel is running pilot programmes in schools in Georgia and Florida, using drones designed to swarm and crash into potential school shooters.
These developments raise serious questions. In chaotic situations, deploying physical force via drone could put bystanders at risk. And once armed drones become normalised in one context, they tend to spread to others. The EFF argues that relying on companies’ internal ethics is not enough — clear laws are needed.
How Europe is handling it
The EU AI Act takes a different approach. It requires high-risk AI systems to undergo conformity assessments before they can be deployed. Autonomous drone systems that make decisions about physical force would likely fall under the highest risk category, meaning they would need to demonstrate safety, transparency, and human oversight before being used.
EASA, the European aviation safety regulator, has identified three areas where AI will play a critical role in drones: detect and avoid systems, adaptive deconfliction, and autonomous navigation. They have also flagged concerns about Level 3B AI — where the system makes non-overridable decisions. These are precisely the areas where regulation needs to be strongest.
Ireland, as an EU member, will follow these regulations. For Irish businesses that use drones — in agriculture, construction, surveying, or logistics — the EU AI Act means you will need to ensure your systems comply with safety and transparency requirements.
What this means for your business
AI regulation is not just about drones. The same principles apply to any AI system that makes decisions affecting people. If you use AI for hiring, credit checks, insurance pricing, or customer assessments, you need to understand your obligations under the AI Act.
The core requirements are: transparency about when AI is being used, human oversight of important decisions, and documentation showing your system works as intended. These are not unreasonable demands, but they do require planning.
Practical steps to take now
Start by inventorying where your business uses AI. Do you use an automated system to screen job applicants? An AI tool to approve credit or set prices? A chatbot that handles customer complaints? Each of these may fall under the AI Act depending on how they are used.
Talk to your software vendors about whether their products comply with EU regulations. If they cannot give you a clear answer, that is a red flag. The companies that take AI regulation seriously will be the ones you can trust with your business data.
AI regulation is coming. The businesses that prepare for it now will avoid costly surprises later.