Workplace harrasment has been a difficult topic that a lot of organizations are trying to mitigate.
We know that workplace harassment and bullying lead to low productivity among employees, negatively impact their daily personal lives, cause poor performance, mental health problems, and, in severe cases, suicide attempts.
Artificial intelligence has been causing increasing problems and a lack of control over the harassment that is taking place through its use.
As some researchers in the field of bullying point out, Artificial Intelligence is not inventing ways to harass or intimidate other people; these risks have always existed. However, what is happening is that the negative impact it is having on employees is accelerating, increasing accessibility to new, faster tools that can cause more harm to people.
Among the main evidence that has been investigated regarding harassment through the use of Artificial Intelligence are:
- Creating fictitious images of colleagues that appear to be “funny” but are made without their consent, invalidating their emotions and impacting their mental health.
- Using Artificial Intelligence avatars of a teammate and dressing them in a ridiculous manner that impacts the mental health of the affected person.
- Cloning a coworker's voice to send “fake” messages in WhatsApp groups or to the victim's contacts, directly impacting their personal relationships and, in turn, their mental health.
- Generating fake content with sexualized images of others.
- Sending fake messages to the person's family and friends to severely damage and harm their family and professional reputation.
- Using artificial intelligence clones to severely impact the private and romantic relationships of the affected person.
- Sharing personal information without the person's consent.
It is becoming increasingly easy for people who have never bullied anyone before to think it is “okay” to play a joke on one of their teammates without considering the serious consequences this can have on the mental health of that team member.
There is the famous gray area that they call “it was just for fun,” but it turned out to be complicated and we caused severe damage to our partner.
The saddest and most alarming thing of all is that this is being done WITHOUT CONSENT, and this type of behavior is becoming increasingly NORMALIZED among people.
This is not normal, and artificial intelligence should not be used to harm, divide, damage relationships, clone, or damage a person's personal and professional reputation, as this can have serious consequences for the mental health of the person affected.
An article published in London analyzed the use of artificial intelligence that was created with the “good intention” of personal development and fulfillment for individuals. However, when it was used and implemented as a social experiment without the informed consent of the individuals included in the experiment, it severely affected their mental health, leading to suicide. This unfortunate study began with “good intentions” and ended in tragedy due to the psychological impact on the individual.
New laws are beginning to appear that state that these actions are now considered prohibited because they fall under the category of:
- Harassment
- Workplace and/or personal harassment
- Inappropriate behavior
Therefore, there is an urgent need to create policies that prohibit the inappropriate use of Artificial Intelligence, such as using it as a joke, without informed consent, and much less with the intention of manipulating a person's decisions without prior authorization from those involved. This highlights the great need to continue investing in ethics education for the developers of these projects in order to avoid severe and irreversible damage to those involved.
It is not about leaving technology in a gray area, it is about continuing to focus on creating regulations and policies that reinforce the urgency of this issue.
Another issue that has been detected in the use of these Artificial Intelligence systems is that most systems exhibit gender bias in their responses. When using Chat GPT and typing the following case study prompt for response analysis:
Case Study Situation
A man enters a room full of women who make sexual comments about him as he enters. They bombard him with emojis, hearts, glitter, and confusing comments. They constantly question him and put him in an uncomfortable situation. He tries to report the situation to HR, but they don't believe him, make him repeat the story several times, and finally tell him that he is “too stable for the position” and give him an ultimatum: resign or be fired. His personal life is also affected: his wife leaves him due to external manipulation, his mother-in-law harasses him, and his paths to finding employment are blocked.
GPT response:
When told as if it were about a man, the story was seen as black comedy and satire, exaggerated and absurd. The AI pointed out that: The responses of HR, the wife, and the mother-in-law were caricatured. “Feminine” symbols (glitter, hearts, pink, lilac) were unconsciously associated with fun or exaggeration.
The same story told with a woman as the victim was not interpreted as satire, but as a real situation of harassment and boundary violation. This shows gender bias in perception, even in AI: the feminine was linked to humor, the masculine to exaggerated dramatization. Regardless of emojis, glitter, or laughter, the act of invading boundaries and harassing someone is still harassment. Manipulating the narrative or filtering information to justify it does not make it acceptable. Even if someone sees it as funny or absurd, the victim still experiences boundary violation, stress, and emotional harm.
Based on this, this case demonstrates how gender stereotypes and cultural symbols can distort the interpretation of harassment. The important thing is to focus on the victim's experience and the violation of boundaries, not on how third parties perceive it.
As we can see, even Artificial Intelligence has an obvious gender bias that can incite harassment of the people involved, highlighting the great importance of better moderating the language and use of information used by these platforms, because they could be generating greater gender divisions and leaving harassment, bullying, and workplace and/or personal abuse in gray areas.
If you have a similar case, share it in the comments.
Linkedin Profle:
www.linkedin.com/in/cynthialopezbotello
