Killer Robots in War and Civil Society - Noel Sharkey #TOA15

Summary

The speaker discusses the dangers of autonomous weapon systems (AWS), also known as 'killer robots,' highlighting their potential to operate without human oversight, violating international humanitarian law. These weapons use advanced technologies like sensors and algorithms to target and kill without human intervention. The speaker critiques their inability to distinguish between combatants and civilians and the unpredictability of their actions. He emphasizes the urgent need for global regulation, including the prohibition of AWS, stressing the ethical and moral implications of machines making life-and-death decisions. The talk also covers the risks of AWS in civilian contexts like policing and border control.

Takeaways

• 😀 The shift in weapon systems: Historically, humans have controlled weapons directly, but modern weapons now often involve computer chips that assist with targeting and firing.
• 😀 Autonomous Weapon Systems (AWS): These are weaponized robots capable of seeking out and eliminating targets without human involvement, raising concerns about their ethical use.
• 😀 Challenges to international law: AWS cannot fully comply with international humanitarian law, especially in terms of distinguishing between combatants and civilians.
• 😀 The concept of 'proportionality' in warfare: AWS cannot reliably assess military advantage or comply with the principle of proportionality, which governs the use of force in warfare.
• 😀 Unpredictability of automated weapons: Machines, particularly swarms of AWS, are unpredictable and could lead to unintended consequences in conflict.
• 😀 Global arms race concerns: If AWS proliferate, it could lead to a new arms race, where many nations possess similar autonomous weapon capabilities.
• 😀 Speed of warfare: The rapid pace of modern warfare, especially with technologies like the Falcon unmanned aircraft, presents challenges for human decision-making.
• 😀 The threat of hacking: AWS are vulnerable to spoofing, trickery, and hacking, which could lead to dangerous and unintended outcomes.
• 😀 Accidental conflicts: The autonomous nature of these weapons means that accidental conflicts could escalate quickly, with no human ability to intervene once the systems are activated.
• 😀 Civilian applications: AWS are already beginning to be used in civilian settings, such as border control or policing, raising concerns about their role in suppressing populations and managing civil unrest.

Q & A

What are lethal autonomous weapon systems (LAWS)?

Lethal Autonomous Weapon Systems (LAWS) are machines or robots that can independently identify and engage targets, including killing them, without human involvement once activated. They utilize sensors and computer algorithms to process information and make decisions.

What is the primary concern with lethal autonomous weapon systems?

The main concern is that these systems cannot fully comply with international humanitarian law, including the Geneva Conventions, which require the ability to discriminate between combatants and civilians and ensure proportionality in warfare decisions.

Why are autonomous weapons considered difficult to control under international law?

Autonomous weapons struggle to reliably distinguish between different types of targets (e.g., combatants vs. civilians) and make nuanced decisions like sparing surrendering soldiers or avoiding harm to civilians in conflict zones. The lack of human oversight makes compliance with international law highly questionable.

What is the issue with 'proportionality' in autonomous weapon systems?

Proportionality in warfare refers to ensuring that the harm caused by an attack is proportionate to the military advantage gained. Autonomous systems cannot accurately assess military advantage or make this judgment, especially in complex or ambiguous situations like the 'fog of war.'

What type of weapon systems are currently in development or use as autonomous weapons?

Examples of autonomous weapons in development or use include the U.S. X-47B, the Israeli Guardian, DARPA's Crusher, and various air-to-air combat drones from China, all of which demonstrate the capability for autonomous targeting and engagement.

How have drones already contributed to the proliferation of autonomous weapon technology?

Drones, which were first used in missile strikes in 2001, have rapidly proliferated, with 87 countries now possessing drones, and 30 of them having armed drones. This demonstrates the fast spread of unmanned systems, which raises concerns about the future of fully autonomous weaponry.

What potential dangers do autonomous weapon systems present in warfare?

Autonomous weapon systems may cause unpredictable outcomes in conflict, including accidental escalations, hacking vulnerabilities, and uncontrollable swarms of weapons, which could result in unintended consequences and harm to civilian populations.

What ethical dilemma is raised by the use of autonomous weapons in warfare?

The ethical dilemma centers on whether we should allow machines to have the authority to kill without human intervention, as it raises serious moral questions about accountability, the value of human life, and the consequences of ceding life-and-death decisions to algorithms.

What role has the United Nations played in addressing the issue of lethal autonomous weapons?

The United Nations has been active in addressing the issue through campaigns and discussions, leading to the establishment of a committee focused on banning such weapons. This includes gaining support from over 80 countries for the need to control or prohibit autonomous weapon systems in warfare.

What are some of the civil uses for autonomous weapon systems, and why is this concerning?

Autonomous weapons are starting to be used for civilian purposes, such as crowd control or border security. For example, drones have been deployed to monitor property, deliver pepper spray, or even fire taser darts at intruders. This civilian application raises concerns about privacy, human rights, and the potential for misuse in oppressive scenarios.