With 1.3 million active personnel, 800,000 reserve personnel, and a budget of $597 billion dollars (3.3% of GDP), the US military delivers the value of protecting the security of the United States. Historically, this task has been accomplished with “troops on the ground,” but as technology improves, robots have begun taking the place of soldiers. Though this transformation has many benefits, there are significant ethical and technological issues to consider.
The operating model of the U.S. Military has evolved over time. In addition to protecting the U.S., the military has been increasingly involved in global security issues (the U.S. military currently has operations in Yemen, Uganda, Iraq, Syria, and Cameroon). Operationally, the U.S. Military is an impressive machine. The organization processes incredible amounts of data to make decisions and is able to effectively manage nearly 2 million people in a highly hierarchical structure.
Ground and aerial robots were first used during war in 2002 in Afghanistan and their use in the U.S. military has since increased dramatically. In 2000, the U.S. military had fewer than 50 unmanned aerial vehicles (UAV), by 2010 they had more than 7000. Most military robots are piloted remotely and are therefore “semi-autonomous” but technology is trending towards fully autonomous vehicles. In 2014, PackBots (the most common military robot) were given some autonomous functionality enabling them to travel to pre-determined waypoints. The Navy’s Phalanax Close-In Weapons System is fully autonomous and is enabled to shoot down anti-ship missiles in a last-resort situation. Piloted robots save lives on the battlefield and are useful in many situations including bomb disposal and building searches. However, as we begin to consider autonomous robots that think for themselves without human intervention, the value proposition gets more complicated.
Autonomous robots have been described as the “third revolution in warfare after gunpowder and nuclear arms.” Unlike nuclear, robots are cheap to build and do not require hard-to-obtain materials – many experts believe the development of autonomous weapons will trigger an arms race. Robots would have to be programmed to make decisions in complex wartime situations and to assess extremely complicated situations (Ex: Is the individual running away from me a civilian or a combatant?). Robotic behavior would never be 100% predictable due to the wide variety and uncertainty of inputs contributing to autonomous decision making. Autonomous weapons lower the threshold of going to battle due to the decrease in loss of life. Furthermore, they remove the human element and disconnect us from the reality of war.
There are arguments to be made in favor of autonomous weapons. Robots do not have emotions influencing decisions. Autonomous weapons may eliminate instances of war crimes. One particularly intriguing example of autonomous robots is the Battlefield Extraction-Assist Robot (BEAR) that was being developed to recognize injured soldiers and transport them away from the battlefield.
The value of human lives is difficult to balance against the potential negatives of fully autonomous robotics. There are several steps that could be taken to protect against potential negative outcomes.
- Stick with semi-autonomous weapons: Semi-autonomous or “piloted” weapons keep a human connected to the robot. Decisions are still made by a human being (with an understanding of the realities of war) but soldiers’ lives are protected. This also creates a clear sense of responsibility. If a mistake is made, the burden falls on the individual. There is no confusion over who should take blame for a robot’s actions.
- Global Community Responsibility: Several large organizations (Future of Life Institute) and notable individuals (Stephen Hawking, Elon Musk) have spoken out against the use of fully autonomous weapons. The global AI community and intergovernmental organizations need to discuss and define international policy on development and use of autonomous weapons. This is crucial to preventing the mentioned AI arms race. Policy and regulation could be used to maintain a high barrier of entry to war by limiting the development of robotic weapons.
- Value Added Applications: The use of fully autonomous robots in some situations such as the BEAR rescue robot mentioned earlier is an opportunity to leverage AI for value adding purposes with far fewer risks of negative consequences. Identifying these opportunities and incentivizing development of these technologies is a good way to embrace the trend of autonomy.
- Targeted Technical Development: If fully autonomous weapons are an inevitable trend in military operations, it is crucial that we tackle the existing technical challenges – notably advanced sensor development and real-time data processing. This technology is a necessity for ensuring robots do not mistakes based on faulty input data.
Digital transformation has changed much about the world and warfare has not been excluded. To ensure the world is secure for many years to come, it is crucial we are intentional about how we approach the use of autonomous robots in war.
 “Department of Defense (DoD) Releases Fiscal Year 2017 President’s Budget Proposal,” press release, February 9, 2016, on Department of Defense website, http://www.defense.gov/News/News-Releases/News-Release-View/Article/652687/department-of-defense-dod-releases-fiscal-year-2017-presidents-budget-proposal, accessed November 2016.
 Lora G. Weiss, “Autonomous Robots in the Fog of War,” IEEE Spectrum, July 27, 2011, http://spectrum.ieee.org/robotics/military-robots/autonomous-robots-in-the-fog-of-war, accessed November 2016.
 Raytheon, “Phalanx Close-In Weapon System,” http://www.raytheon.com/capabilities/products/phalanx/, accessed November 2016.
 Future of Life Institute, “Autonomous Weapons: An Open Letter from AI & Robotics Researchers,” July 28, 2015, http://futureoflife.org/open-letter-autonomous-weapons, accessed November 2016
 Barb Ruppert, “Robots to rescue wounded on battlefield,” November 22, 2010, https://www.army.mil/article/48456/robots-to-rescue-wounded-on-battlefield/, accessed November 2016
 Jon Cartwright, “Rise of the robots and the future of war,” November 20, 2010, https://www.theguardian.com/technology/2010/nov/21/military-robots-autonomous-machines, accessed November 2016