Guidelines for Military AI Use from Tech Executives

The Defense Innovation Board has handed down a series of guidelines the advise the military on how AI in military conflict should operate. The initiative comes from an understanding that precautions should be taken to minimize damage to non-combatants and their property. The Defense Innovation Board has been in existence since 2016. It was developed to enable easier collaboration between Silicon Valley and the US Military. Among the board members are current and past executives of Big Tech giants Google, Facebook, and Microsoft.

Incorporating AI Tech into Military

With the resurgence of building AI into many systems, an interest to develop its use in military technology is an obvious step. The development of military-grade AI carries with it several moral and ethical underpinnings. The Defense Innovation Board has decided to outline principles that should be applied to any military decisions regarding AI. The intention is to create a best-practice regarding the use of artificial intelligence on the battlefield to limit its contrary abuses.

Five Principles of Operation

The Defense Innovation Board has set out five ethical perspectives in which military AI decisions should be framed: Responsible, Governable, Reliable, Traceable, and Equable. Using these principles as a basis, the board suggests twelve recommendations to aid the US military in incorporating AI into their operations while still maintaining ethical transparency. The report presented by the committee advises the military to create a risk management methodology that can be used to gauge potential adverse outcomes of the program.

The innovation of AI in military tech is already underway, as the US Department of Defense announced that it was developing a new missile system using AI as an aid to guidance. In 2018, Google failed to renew their interest in a project codenamed Maven, which saw the company using AI to determine potential military targets. The incorporation of AI into military technology isn’t slowing down, and the board’s recommendations are the best option the military has to build accountability into their AI systems so far. Only time will tell if they will utilize it effectively.