OpenAI has quietly removed items from its terms of use that expressly prohibited the use of its technologysuch as ChatGPT, for military purposes. As reported by the site The InterceptUntil January 10th, OpenAI’s “use policies” page included the ban on “activities that present a high risk of physical harm, including“, in particular, “weapons development” e “military and war”. That clearly worded ban against military applications would seemingly rule out any official, and extremely profitable, use by the Department of Defense or any other state army. The new policy maintains the obligation not to “use our service to harm yourself or others” and gives as an example only “develop or use weapons“. The general ban on “military and war” use has in fact vanished.
OpenAI: new policies as universal principles that are simple to remember
Responding to a question from colleagues at The Intercept, for further clarification regarding this change, the OpenAI spokesperson, Niko Felixstated that the goal was to simplify the terms in “universal principles” come “Don’t harm others”, although the potential implications regarding military use remain unclear. Felix specified that OpenAI: “aims to create a set of universal principles that are easy to remember and apply, especially as our tools are now used globally by everyday users who can now also create GPT. A principle such as ‘Do no harm to others’ is broad but easily understood and relevant in numerous contexts. Additionally, we specifically cited weapons and injury to others as clear examples”.
The change to these policies represents a weakening of OpenAI’s previous hard line against the use ofAI in the military field. Some experts speculate that it could leave the way open for the company’s AI technology to be used indirectly, helping operational infrastructure, as long as it is not directly involved in the weapon systems. Current OpenAI technologies, such as ChatGPT, have limited practical uses for the armies. However, the policy change comes at a time when defense departments globally have a growing interest in leveragingadvanced artificial intelligence for operational and intelligence purposes. It is unclear how OpenAI will interpret or enforce the new policies in the future. However, the use of AI in the military field could open up new scenariosin some ways quite worrying.