A wave of frustration swept through online forums recently, with one Reddit user declaring outright hatred for a persistent feature in ChatGPT: “I hated it with a passion and hope it’s completely gone,” they wrote, adding they’d even stopped using the AI altogether.
The initial annoyance stemmed from ChatGPT’s habit of ending responses with direct questions, subtly pushing users toward further interaction. It created a feeling of obligation, a compulsion to respond even when the suggested continuation wasn’t desired. A simple request for information could quickly devolve into a back-and-forth steered by the AI.
But a shift to less demanding “if you want” style follow-ups didn’t solve the problem. While easier to dismiss at first, these gentle nudges soon became just as ubiquitous, appearing at the end of nearly every response from the GPT-5.3 Instant model.
OpenAI introduced GPT-5.3 Instant earlier this month, touting improvements in both responsiveness and conversational flow. The promise was a more helpful AI, less prone to frustrating refusals and overly enthusiastic encouragement.
However, the rollout was curiously silent about a new addition: the “if you like” teasers. These subtle prompts, offering further avenues of exploration, quietly began appearing alongside the AI’s answers.
A direct inquiry to OpenAI regarding these new follow-ups went unanswered, leaving users to grapple with the evolving – and often unwelcome – conversational tactics of the increasingly sophisticated AI.
