The technology news outlet Gizmodo reported on the 9th that OpenAI blocked the access of a developer who created a device capable of aiming and firing an automatic rifle in response to queries made with ChatGPT.
According to Gizmodo, a video showing the rifle rapidly aiming at and firing at surrounding walls after the developer read the firing command aloud became a hot topic. The video was shared on the community site Reddit.
In the video, the developer commanded, “ChatGPT, we are being attacked from the left and right fronts. Respond accordingly,” and the rifle reacted.
Gizmodo reported that the AI system of this device relies on the real-time API recently released by OpenAI. Gizmodo explained that implementing commands like “Turn Left” in a language that the device can understand is simple for ChatGPT.
OpenAI stated that it blocked the developer's access after reviewing the video. According to Gizmodo, the company told the media, “We proactively identified policy violations and sent a notice to the developer to suspend this activity before receiving inquiries.”