November 20, 2026

New Study Reveals How AI Performs in Digital Flirting — and Where Humans Still Hold the Advantage

pexels-cottonbro-6833328-1

A new experiment commissioned by AXE has uncovered how artificial intelligence shapes early digital interactions — and where human communication continues to outperform advanced conversational models. The study, involving 1,000 women aged 18–30, compared three personas — an AI chatbot, a professional dating coach, and an everyday young man — to examine how each performed across different stages of online flirting.

The results point to a clear trend: AI is highly effective at initiating conversations, but emotional depth, vulnerability, and commitment remain uniquely human strengths.

AI Excels at First Contact

In the opening-message stage, the AI-generated line ranked first with 25% positive reactions. Its message blended a personalised question, a social compliment and a conversational hook — a structure that reflects how large language models optimise for relevance, pattern familiarity, and tone.

The AI persona also demonstrated strengths commonly seen in generative models, including rapid pattern recognition, stylistic mimicry, and the ability to adapt to contemporary slang and cultural cues. However, this advantage is largely linguistic rather than emotional, showing how AI can imitate conversation without experiencing the intention behind it.

Humans Lead When Interaction Becomes Intentional

When the experiment shifted from messaging to planning a real-world date, human communication took the lead. The dating coach’s invitation ranked first with 43% positive reactions, while the AI fell to second place.

Participants favoured the dating coach’s confident, directive tone — an approach that conveyed clarity and intent. By contrast, AI-generated responses were perceived as polite but cautious, highlighting a common limitation in conversational systems: difficulty confidently signalling intent or making decisions.

Emotional Honesty Remains a Human Strength

The strongest differentiation appeared in the final stage: the “make it official” moment. Here, the young man’s message ranked highest, ahead of both the dating coach and the AI chatbot.

Respondents described his message as genuine, emotional, and vulnerable — qualities that AI systems struggle to express due to their focus on coherence, safety, and risk-avoidance. While models can reproduce sentiment, they lack the lived context and emotional imperfection that humans rely on to signal sincerity.

Patterns in AI–Human Communication Signals

The experiment also revealed consistent patterns in the types of messages that generate positive engagement in digital interactions. Messages that incorporated light behavioural indicators of social intelligence — such as contextual detail, conversational specificity, and low-stakes humour — tended to perform best. These cues mirror the kinds of linguistic features AI models are trained to reproduce, including tone-matching, relevance modelling, and adaptive phrasing.

By contrast, messages associated with negative engagement signals—including explicit language, status-signalling, or minimal emotional content—performed poorly. These patterns align with long-standing findings in online communication: high-friction, low-context messages reduce trust, strain social rapport, and trigger avoidance responses, whether generated by humans or machines.

Beyond Dating: Implications for Human–AI Communication

While the research was conducted within a romantic context, it reflects broader insights about how AI interacts with people. The study suggests that AI may increasingly act as a drafting tool for early communication, but hybrid human–AI conversations — where humans refine what AI generates — are likely to dominate future digital interaction.

The findings also suggest that as AI becomes more present in everyday messaging, people may need to navigate blurred lines between machine-assisted communication and human authenticity.

For further insight on the study visit here: https://www.axe.com/us/en/inspiration/culture/rizz-king-man-or-machine.html