Why do people so often misread each other's signals during sex?
The core problem is a systematic mismatch between how people express consent or refusal and how their partners interpret those signals. A 2022 study of 150 adolescents and young adults (ages 14-24) found stark gender differences: 95% of sexually active females said they used verbal cues to communicate refusal, but only 57% of males interpreted those same verbal cues as refusal [1]. In other words, nearly half the young men in the study were missing a clear verbal 'no' from a partner. The same study found that about half of both males and females interpreted a partner's silence or lack of response as consent [1]. This means that indirect or ambiguous language—or the absence of any clear language—creates a dangerous gray zone where one person believes they have permission while the other does not.
This gap isn't just about gender. A study of sexual and gender minority (SGM) college students found that while verbal communication was the most common way to signal consent, it fell on a spectrum from a gentle request ('Can I?') to a command ('Do this') [2]. The same spectrum applied to refusals, from a soft 'no' to a firm 'stop.' The researchers noted that the meaning of these signals can shift depending on tone, context, and relationship history, making it easy for partners to interpret the same phrase differently [2]. When language is vague or indirect, the risk of misunderstanding—and therefore of non-consensual contact—goes up.
How does pop culture shape the way we talk (or don't talk) about consent?
The language people use during sex isn't invented in a vacuum—it's heavily influenced by the media they consume. A 2023 analysis of nearly 600 popular songs from pop, hip-hop, and country genres found that the most common form of sexual communication in lyrics was a sexual demand (e.g., 'Strip that down'), appearing in about one-third of all singing parts [3]. In contrast, direct sexual requests (e.g., 'Can I hit it?') were among the least common [3]. Even more troubling, references to 'inferred consent'—assuming a partner wants sex without asking—appeared in roughly one in five singing parts (e.g., 'I can tell you wanna f***') [3]. These lyrics normalize the idea that you don't need to ask; you can just read your partner's mind or assume their desire. When this script is repeated across hundreds of popular songs, it reinforces the very indirect communication style that leads to the misunderstandings documented in the adolescent study [1].
The same analysis found that sexual requests were significantly more common in songs released before the #MeToo movement (2016-2017) than after (2018-2019), suggesting that cultural conversations can shift the language of consent in media [3]. However, the persistence of demands and inferred consent in post-#MeToo lyrics shows that problematic scripts remain widespread. This matters because song lyrics don't just reflect culture—they help shape the sexual scripts that people, especially young people, draw on during real encounters.
If even AI can't reliably read between the lines, how can humans?
A 2025 study tested four different large language models (like ChatGPT) on their ability to judge whether a sexual scenario was consensual based on the language used [4]. When the communication was clear and direct—like an explicit 'yes' or 'no'—the AI models agreed with human experts 100% of the time [4]. But when scenarios involved ambiguous language or intoxication, the AI models frequently disagreed with experts, often misjudging whether consent was present [4]. This is a powerful demonstration that even the most sophisticated language-processing systems struggle with the subtle, indirect, and context-dependent communication that humans commonly use during sex. If an AI trained on billions of words can't reliably interpret a vague 'I'm not sure' or a passive 'whatever you want,' it's unrealistic to expect a partner in the heat of the moment to do so accurately.
Another 2024 study looked at Snapchat's My AI, which is designed to answer young people's questions about sex. It consistently advised users to use 'clear, honest consent and refusal communication through physical contact or direct conversation' [5]. This advice is sound, but the study also noted that the AI's responses varied across users, which could confuse someone seeking consistent guidance [5]. The fact that even purpose-built AI tools give variable advice on consent language underscores how complex and context-dependent this communication really is. The takeaway for real people is simple: the safest, most reliable way to ensure consent is to use direct, unambiguous language—ask clearly and listen for a clear answer.
About These Sources
This answer is built on 5 peer-reviewed studies — published from 2021 to 2025, 2 from 2024 or later, 5 in Q1 journals — selected as the most relevant from 6 studies that passed quality screening, drawn from 47 papers retrieved from a database of over 500 million.
Sources used in this answer
Communication and Interpretation of Sexual Consent and Refusal in Adolescents and Young Adults
In a survey of 150 adolescents and young adults (ages 14-24), females were significantly more likely than males to use and interpret verbal cues for both consent and refusal; for example, 95% of females used verbal refusal cues but only 57% of males interpreted them as refusal, and about half of both genders interpreted silence as consent [1].
Sexual Consent Communication among Sexual and Gender Minority College Students
Among 81 sexual and gender minority college students, verbal communication was the most salient indicator of both consent and non-consent, but it fell on a spectrum from requests to commands, with similar ambiguity in physical cues [2].
“I Can Tell You Wanna F***”: A Content Analysis of Sexual Communication in Popular Song Lyrics
In a content analysis of 584 popular songs (pop, hip-hop, country, 2016-2019), sexual demands and inferred consent (e.g., 'I can tell you wanna f***') were common, appearing in about one-third and one-fifth of singing parts respectively, while direct sexual requests were rare and more common before #MeToo [3].
Exploring How Large Language Models Understand Sexual Communication in Hypothetical Sexual Situations
Four large language models (LLMs) correctly identified consent in scenarios with clear verbal communication, matching expert judgment, but frequently disagreed with experts when scenarios involved ambiguous language or intoxication, showing that AI struggles with nuanced sexual communication [4].
Large Language Models in an App: Conducting a Qualitative Synthetic Data Analysis of How Snapchat’s “My AI” Responds to Questions About Sexual Consent, Sexual Refusals, Sexual Assault, and Sexting
Snapchat's My AI consistently advised users to use clear, direct communication for consent and refusal, but its responses varied across users, and it defaulted to recommending consultation with a trusted adult for sensitive topics like assault [5].