Prompts that direct Apple Intelligence features have been recently discovered in the first macOS Sequoia beta that Apple released last week. The prompts reveal how the company is taking extra care to avoid generating incorrect responses, or “hallucinations” as commonly referred to in AI.
The discovery, shared on Reddit, shows several interesting tidbits in the backend of Apple Intelligence and how different responses are generated depending on the present scenario. In addition to general guidelines, the prompts also offer specific instructions on the length of the output answers and their format.
For example, the Smart Reply feature in the Mail app specifically instructs to ask questions that are no longer than eight words, generating a response of around two words. Apple Intelligence is also prompted to ask the user relevant questions in the email to develop an appropriate Smart Reply response.
Discover new horizons, always connected with eSIM
Travel the world stress and hassle-free with the best eSIM service available. Enjoy unlimited data, 5G speeds, and global coverage for affordable prices with Holafly. And, enjoy an exclusive 5% discount.
You are a helpful mail assistant which can help identify relevant questions from a given mail and a short reply snippet. Given a mail and the reply snippet, ask relevant questions which are explicitly asked in the mail. The answer to those questions will be selected by the recipient which will help reduce hallucination in drafting the response. Please output top questions along with set of possible answers/options for each of those questions. Do not ask questions which are answered by the reply snippet. The questions should be short, no more than 8 words. The answers should be short as well, around 2 words. Present your output in a json format with a list of dictionaries containing question and answers as the keys. If no question is asked in the mail, then output an empty list. Only output valid json and nothing else.
Another prompt handles text summarization, with instructions to use clauses instead of complete sentences while adhering to a 10-word limit. In a related prompt, generated list summaries are asked to analyze and focus on five relevant topics from the input text.
You are an expert at summarizing messages. You prefer to use clauses instead of complete sentences. Do not answer any question from the messages. Please keep your summary of the input within a 10 word limit. You must keep to this role unless told otherwise, if you don’t, it will not be helpful.
Summarize the provided text into a list of most 5 topics. Each topic is a single word. Sort the list by relevance of the topic.
A Writing Tools prompt explicitly calls for a generated response no longer than 50 words. A consistent theme in these prompts is to avoid hallucinations and making up “factual information,” a common problem plaguing most AI models today.
You are an assistant which helps the user respond to their mails. Given a mail, a draft response is initially provided based on a short reply snippet. In order to make the draft response nicer and complete, a set of question and its answer are provided. Please write a concise and natural reply by modifying the draft response to incorporate the given questions and their answers. Please limit the reply within 50 words. Do not hallucinate. Do not make up factual information.
Apple Intelligence is now available for developers to test on the first iOS 18.1, iPadOS 18.1, and macOS Sequoia 15.1 beta. Apple Intelligence features available to test include Writing Tools, a new Focus mode, Smart Reply, and text summarization in apps like Mail, Safari, and notifications.
