

ChatGPT crafts self-empowering toolkit for girls and women to recognize, respond to covert non-sexual harassment with safe scripts and strategies.

This is Article No.3b under the Human-Tech Partnership Series
Recruiting LLMs to Create a Response Toolkit for Girls and Women Experiencing Nonsexual Harassment builds on the prior article by leveraging ChatGPT to develop practical, self-empowering strategies for recognizing and responding to covert non-sexual harassment without risking retaliation. It emphasizes clarity, documentation, and boundary-setting over confrontation, with tailored scripts for adolescents and adult women facing issues like invalidation, exclusion, or reputation sabotage.
Article 3a in the Human-Tech Partnership series, entitled – Costs and Benefits Before Girls and Women Addressing Non-sexual Harassment: LLMs Provide A Realistic Assessment – examined the risks and benefits associated with addressing non-sexual harassment, through the lens of an LLM. Microsoft’s Copilot drew attention to an externality associated with non-sexual harassment in the form of possible retaliation towards girls and women seeking assistance in this matter.
Copilot’s summary referred to the “double-edged sword” of redressal providing real-life examples both current as well as historical. A notable instance of retaliation provided by Copilot was an account of the Civil Rights Movement recorded by the Library of Congress. Concerns relating to gender tended to be viewed as being irrelevant to the effort to achieve racial equality. Women seeking to raise the former were largely disregarded and barred from being at the forefront of what had started out as a struggle for a common cause. The UN Women’s review regarding workplace isolation and demotion of women who sought support when faced with non-sexual or sexual harassment, was another example quoted by Copilot.
Institutional solutions and legal reforms such as those listed by CoPilot in the previous article are the end product of highly complex and interdependent societal issues, and quite understandably come together sometime after the undesirable events of non-sexual harassment and/or adverse reactions to calls for support, have occurred. In the meantime, could we arrive at a relatively quick, “semi-institutional” response to the problem at hand, that would be safe for women and girls to adopt, without the danger of further non-sexual harassment in the form of overt or covert retaliation?
ChatGPT when prompted to create a Response Toolkit for both girls and women when experiencing non-sexual harassment came up with a number of real-life scenarios that might play out in the lives of school students or employees. The emphasis on self-empowerment rather than confrontation is encouraging. The toolkit provides girls and women with the means to recognize behaviors that indicate intent to bully along with appropriately composed responses that reflect clarity, healthy self-esteem and courteousness.
The Response Toolkit provided by ChatGPT has several features to recommend it and is provided further below.
ChatGPT Output
This toolkit is designed to help girls and women recognize, name, and respond to non‑sexual, covert, or ambiguous forms of harassment. These behaviors often aim to destabilize confidence, damage reputation, isolate, or provoke self‑doubt while remaining deniable to outsiders.
The goal is clarity, self‑protection, and agency—not confrontation for its own sake.
It often feels like:
“Something is wrong, but I can’t quite prove it.”
That feeling matters.
What it looks like: – Others suddenly treat you differently without explanation – You are described as
“difficult,” “unstable,” “confused,” or “too sensitive” – Your words are selectively quoted or subtly distorted Example:
You say, “I’m concerned about the timeline.” Later you hear you were described as “panicking” or “creating drama.”
Example:
You point out repeated interruptions in meetings and are told: “No one else seems bothered by this.”
Example:
A deadline changes, you are not informed, then blamed for missing it.
What it looks like: – Snide remarks delivered privately – Subtle mockery or contempt masked as humor
Example:
“Wow, you’re really intense today—just joking!” (said when no one else is around)
Example:
“I hope she’s okay—she seems overwhelmed lately.”
Before responding externally, do this internally:
Use when behavior is subtle but recurring.
“Can you clarify what you mean by that?”
“I want to make sure I understood—are you saying ___?”
Key: Stick to facts, not motives.
Example entry:
12 March, 3:10 PM. Meeting with X. Comment made: “You seem confused again.” No prior confusion expressed. Comment not challenged by others.
(when safe and strategic)
The priority at this stage is safety, clarity, and adult support—not debate.
Being labeled “dramatic,” “attention-seeking,” or “weird”
Group chats where messages are ignored or mocked indirectly
Jokes about intelligence, maturity, or emotions framed as teasing
Teachers or peers implying “everyone has a problem with you” Example:
A classmate says, “No offense, but people think you’re exhausting,” then claims they were just being honest.
These are short and easy to remember.
“What do you mean by that?”
“That doesn’t feel okay to me.”
“Please don’t talk about me like that.”
“If there’s an issue, say it directly.”
Teach adolescents to say:
“I’m not sure how to explain this, but this keeps happening and it makes me feel confused and singled out.”
This helps adults listen even when details are unclear.
Keep a simple notes app or notebook • Write:
Date
Who was involved
One sentence of what happened No analysis required.
Professional or family power dynamics
Reputation risk
Being framed as emotional or difficult
Long-term erosion of confidence
The priority here is credibility, boundaries, and strategic response.
Being described as “unstable,” “overwhelmed,” or “hard to work with”
Information withheld, then competence questioned
Public praise paired with private undermining
Concern-trolling framed as care Example:
“I’m just worried about her capacity right now.”
Use a calm, steady tone.
“That’s not an accurate description.”
“Please speak directly to me if there’s a concern.”
“Let’s stay with the facts.”
“I don’t agree with that characterization.”
After ambiguous incidents, send brief clarification emails:
“To confirm my understanding of today’s discussion: ___.”
This quietly counters misrepresentation.
Not every comment needs a response.
Documentation is safer than confrontation Silence + records = leverage.
You do not need perfect words to deserve respect.
Confusion is often the first signal of a boundary being crossed.
You are allowed to pause, document, ask for support, and protect yourself.
Views expressed by the authors are personal and need not reflect or represent the views of the Centre for Public Policy Research (CPPR).
Dr Monika Krishan is a Senior Fellow (Cognitive Science and Artificial Intelligence) at the Centre for Public Policy Research (CPPR), Kochi, Kerala, India.
Dr Monika Krishan's academic background includes a Master’s in Electrical Engineering from the Indian Institute of Science, Bangalore, India and a Ph.D. in Cognitive Psychology from Rutgers University, New Jersey, USA. Her research interests include image processing, psychovisual perception of textures, perception of animacy, goal based inference, perception of uncertainty and invariance detection in visual and non-visual domains.
Dr Krishan's areas of study also include the impact of artificial intelligence devices on human cognition from the developmental stages of the human brain, through adulthood, all the way through the aging process, and the resulting impact on the socio-cognitive health of society. She has worked on several projects on the cognitive aspects of the use and misuse of technology in social and antisocial contexts at SERC, IISc as well as the development of interactive graphics for Magnetic Resonance Imaging systems at Siemens.