🤖 AI

Gender Differences in AI Literacy Workshop Outcomes and Deepfake Engagement

This study of Australian secondary students reveals that while male students initially reported higher STEM interest and were more likely to create deepfakes, a single-day AI literacy workshop significantly improved knowledge for all participants and notably boosted female students' confidence and career aspirations, thereby partially narrowing the gender gap in STEM.

Original authors: Jake Renzella, Christian Bergh, Natasha Banks, Alexandra Vassar

Published 2026-06-16

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Jake Renzella, Christian Bergh, Natasha Banks, Alexandra Vassar

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a one-day "AI Bootcamp" for high school students in Australia. The organizers wanted to know: Do boys and girls walk into this room with the same skills, and do they leave with the same new superpowers?

Here is the story of what they found, told without the jargon.

The Starting Line: Who Thought They Were the Experts?

Before the workshop started, the researchers took a "snapshot" of the students' minds.

The Boys' Confidence: The boys generally felt like they knew more about how AI works. It's like a group of kids where the boys are more likely to say, "I know how to fix the engine," even if they haven't actually opened the hood yet. They were also much more interested in becoming engineers or computer scientists later in life.
The Girls' Habits: The girls, however, were the ones actually using the tools. They were more likely to ask AI for help with homework or to get advice on a problem. If AI were a library, the boys were the ones bragging about knowing the Dewey Decimal System, while the girls were the ones actually checking out books to write their essays.
The "Deepfake" Game: This is where things got interesting. A "deepfake" is like a digital puppet show where AI makes a video of someone saying or doing things they never actually did.
- Boys were much more likely to have shared these fake videos with friends.
- Girls were more likely to have reported (told an adult or platform) that they saw a fake video.
- The Metaphor: Imagine a playground. The boys were more likely to be the ones passing around a funny but fake rumor, while the girls were the ones raising their hands to tell the teacher, "Hey, that rumor isn't true."

The Workshop: The "Magic Lesson"

The students spent one day learning about AI. Afterward, the researchers took another snapshot.

Everyone Got Smarter: Both boys and girls learned a lot. Their knowledge about AI jumped up significantly, like filling a bucket with water. They both felt more confident, too.
The Girls' "Super-Boost": While everyone learned, the girls got a broader upgrade.
- They didn't just learn facts; they understood the concepts better (like how AI is trained or how it can make mistakes).
- The Big Win: The girls' interest in becoming AI or computer science experts grew significantly. It was like a spark catching fire.
- The Boys' Stasis: The boys' interest in these careers didn't really change. They were already interested, but the workshop didn't make them more interested.
- The Result: Because the girls' interest went up and the boys' stayed the same, the "gap" between them got a little smaller. The workshop helped level the playing field.

The Deepfake Safety Lesson

The study found a clear pattern in how students handled "fake" content. Since the boys were more likely to share deepfakes, the researchers suggest that future lessons need a special "safety module" for them. It's like teaching a specific traffic rule to the drivers who are most likely to speed, rather than just giving a general lecture to everyone.

The Bottom Line

This study is like a mirror held up to a classroom. It shows that:

Boys often feel more confident about AI but are less likely to use it for schoolwork and more likely to share fake content.
Girls are active users of AI for learning and are more cautious about fake content.
One Day Can Matter: A single-day workshop didn't just teach facts; it helped girls feel more excited about future careers in tech, helping to close the gap between boys and girls in the tech world.

Important Note: The study only looked at two schools in regional Australia, so we can't say this happens everywhere, but it gives us a strong clue about how to design better AI classes for everyone.

Technical Summary: Gender Differences in AI Literacy Workshop Outcomes and Deepfake Engagement

Problem Statement
As Artificial Intelligence (AI) literacy initiatives expand within K–12 education, there is a critical gap in understanding how gender influences student baseline perceptions, tool usage, and responsiveness to interventions. While gender disparities in computing education are well-documented, empirical evidence regarding gender differences in AI literacy among secondary students remains sparse. Furthermore, the rapid rise of synthetic media (deepfakes) introduces gender-differentiated safety risks, yet the intersection of AI literacy education and deepfake awareness has received limited empirical attention. Existing frameworks often address what students should know but fail to analyze who benefits from these interventions or whether strategies are effective across diverse demographic groups.

Methodology
This study evaluated the impact of a one-day AI literacy workshop (part of the "Day of AI" program) on Australian secondary students from two co-educational government schools in New South Wales.

Participants: The study included students from Years 7, 8, and 10 ( $N_{pre} = 199$ , $n_{post} = 136$ ). Single-sex schools and Year 9 were excluded to avoid confounding variables. Non-binary or non-disclosed gender responses were excluded due to insufficient sample size ( $n \le 9$ ).
Instrument: A survey adapted from prior work measured nine constructs: AI knowledge, understanding of AI training and bias, attitudes, confidence, tool usage frequency, conceptual application, motivations, deepfake experience (multi-select), and STEM career interest.
Analysis: Statistical analyses controlled for year level and school.
- Ordinal items (Likert scales): Analyzed using Ordinary Least Squares (OLS) regression with gender dummy-coded (Male as reference). Effect sizes were calculated using Cohen's $d$ .
- Binary/Multi-select items: Analyzed using logistic regression, reporting Odds Ratios (OR) for the gender coefficient.
- Intervention Effects: Pre- and post-intervention comparisons were conducted within gender groups using a model outcome of $\sim phase + year\_level + school$ .

Key Findings

1. Baseline Gender Differences (Pre-Workshop)

STEM Aspirations: Male students reported significantly higher interest in STEM careers across all three domains: Engineering ( $d = -0.81$ ), Computer Science ( $d = -0.68$ ), and AI ( $d = -0.58$ ).
AI Understanding: Males reported significantly higher self-assessed understanding of how AI systems are trained ( $d = -0.29$ ).
Tool Usage: Female students were significantly more likely to use AI for schoolwork (OR = 1.81) and to seek advice from AI tools (OR = 2.35). Males reported more frequent use of Microsoft Copilot.
Deepfake Behaviors: Males were significantly more likely to have shared deepfake content (23.2% vs. 10.0%, OR = 0.34). Females were significantly more likely to have reported deepfake content (24.0% vs. 14.1%, OR = 2.14). Trends suggested males were more likely to create deepfakes or be subjects of them, though these were not statistically significant after controlling for covariates.
No Significant Differences: Baseline differences were not found in general AI knowledge, confidence, privacy awareness, or understanding of AI bias.

2. Differential Intervention Effects (Post-Workshop)

General Gains: Both genders showed statistically significant improvements in AI knowledge and confidence (medium effect sizes for both).
Female-Specific Gains: Female students demonstrated a richer profile of gains, including medium-sized improvements in understanding AI training ( $d = 0.75$ ) and AI bias ( $d = 0.49$ ). Crucially, females showed significant increases in career interest for AI ( $\Delta M = +0.45$ , $d = 0.52$ ) and Computer Science ( $\Delta M = +0.44$ , $d = 0.47$ ).
Male-Specific Outcomes: Male students showed no significant change in career aspirations for AI or Computer Science.
Platform Recognition: Both genders improved in identifying AI usage in entertainment platforms (Netflix, Spotify), with females also showing gains in identifying TikTok.

Key Contributions and Claims of Significance
The authors claim to provide the first gender-disaggregated analysis of an AI literacy workshop with Australian secondary students. The study's significance lies in three primary areas:

Narrowing the Gender-STEM Gap: The workshop appears to help narrow the gender gap in STEM aspirations. While males started with higher interest, their interest remained static post-intervention, whereas females showed significant gains in AI and CS career interest.
Gender-Responsive Outcomes: The intervention produced gender-differentiated outcomes. While both genders gained in knowledge and confidence, females achieved broader conceptual gains (training, bias) and shifts in career trajectory.
Deepfake Safety Implications: The findings highlight distinct behavioral patterns regarding synthetic media. The higher rate of sharing and potential creation of deepfakes among males suggests a need for targeted safety education focusing on the ethical dissemination of synthetic media for male students.

Limitations and Recommendations
The authors note limitations including the restriction to two regional schools (limiting generalizability), the anonymous survey design preventing paired individual analyses, and potential self-reporting bias.

Based on these findings, the paper proposes three specific recommendations for AI literacy program design:

Leverage workshops specifically to increase female STEM career interest.
Incorporate gender-specific deepfake safety modules, particularly targeting male students.
Ensure workshop evaluations explicitly monitor the equity of impact by gender.

The Starting Line: Who Thought They Were the Experts?

The Workshop: The "Magic Lesson"

The Deepfake Safety Lesson

The Bottom Line

Technical Summary: Gender Differences in AI Literacy Workshop Outcomes and Deepfake Engagement

More like this