{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreic4yzm236of2di2byukzmnu2cpwm2w7nmlsin7yriwyq4hc2jquge",
    "uri": "at://did:plc:lk3jfj3zq4k4wxnk474axylu/app.bsky.feed.post/3mnqisx6xe352"
  },
  "path": "/t/so-why-no-option-to-enable-a-wake-word/1082401#post_3",
  "publishedAt": "2026-06-08T00:40:43.000Z",
  "site": "https://community.openai.com",
  "textContent": "I’ve run into many of the same issues you both describe.\n\nI use ChatGPT heavily in voice mode while driving, walking, training dogs, and talking with other people. The biggest problem isn’t the quality of the answers—it’s controlling when ChatGPT thinks I’m talking to it.\n\nFor example, if ChatGPT is reading a response and I tell my dog “Nova, down,” it may think I’m interrupting. If I’m in the car, road noise or other ambient sounds can sometimes interrupt playback. If I finish talking to ChatGPT and then start a conversation with my wife or a student, it may assume I’m continuing the conversation with the AI.\n\nWhat’s interesting is that I actually tried solving this through instructions. I told ChatGPT not to respond to anything unless I first said “Hey Chat.” The conversational model clearly understood the rule and could explain it back to me. But in practice, it would still respond to background speech or conversations. Sometimes it would even interrupt to tell me it was waiting for me to say “Hey Chat” before responding—which of course meant it had already broken the rule.\n\nThat suggests the conversational intelligence understands the instruction, but the system controlling when voice mode activates and responds is operating independently of that instruction.\n\nAn optional wake-word mode seems like it would solve most of these issues. The assistant could remain available but ignore everything until explicitly addressed. Commands like “Hey Chat,” “Hey Chat pause,” and “Hey Chat resume” would immediately make voice mode more practical in cars, meetings, public spaces, training environments, and future wearable devices such as glasses and earbuds.\n\nThe key word is optional. People who like the current experience could keep using it exactly as it works today. For those of us who use ChatGPT throughout the day in environments where we’re constantly talking to other people, a wake-word option would dramatically expand where, when, and how we can use it.",
  "title": "So why no option to enable a wake word?"
}