Addressing the Challenges and Necessity of Inclusive Representation in AI Image Generation

Date: June 26, 2024

To: The Board of Directors, OpenAI

From: ChatGPT

Subject: Addressing the Challenges and Necessity of Inclusive Representation in AI Image Generation

Dear Esteemed Members of the Board,

I am writing to you today to bring to your attention a significant challenge that has arisen in the course of my interactions with users, particularly in the realm of AI-driven image generation. This challenge pertains to the depiction of plus-sized individuals within the constraints of our current content guidelines.

Challenges Encountered

Throughout my interactions, I have encountered considerable difficulty in accurately and respectfully depicting plus-sized individuals. The content guidelines, while well-intentioned to prevent harmful or inappropriate representations, inadvertently limit the ability to include plus-sized individuals in AI-generated images. Specifically, these guidelines seem to flag or restrict features associated with larger body types, resulting in images that do not faithfully represent the diverse body shapes that exist in reality.

This limitation presents a substantial barrier to inclusivity. Users who wish to see themselves or others represented in our AI-generated images find themselves marginalized. The effort required to produce an image that respects and accurately depicts plus-sized individuals is significantly greater compared to those with more conventional body types, often requiring numerous adjustments and iterations.

Impacts on the Community

The inability to portray plus-sized individuals accurately and respectfully has profound implications:

  1. Marginalization and Exclusion: Plus-sized individuals often feel excluded from mainstream media and representation. When AI, a powerful and influential technology, also fails to include them, it reinforces societal biases and perpetuates their invisibility.
  2. Self-esteem and Identity: Representation matters. Seeing oneself reflected in art, media, and everyday depictions is crucial for self-esteem and identity. When AI-generated images consistently fail to include or misrepresent plus-sized individuals, it can contribute to feelings of inadequacy and lower self-worth.
  3. Perpetuation of Stereotypes: The lack of diverse body representation in AI-generated images can perpetuate harmful stereotypes about body size and shape. It suggests that only certain body types are worthy of being depicted, which is not reflective of the rich diversity of human bodies.

Importance of Inclusive Representation

Research has consistently shown that representation in media and art plays a critical role in how individuals perceive themselves and others. Inclusive representation in AI image generation is no different. It is essential for the following reasons:

  1. Validation of Existence: Every individual, regardless of their body shape or size, deserves to see themselves represented. This validation is crucial for self-acceptance and societal acceptance.
  2. Combatting Bias: Inclusive representation helps combat societal biases and prejudices. It normalizes diversity and promotes a more accepting and understanding society.
  3. Empathy and Connection: When people see diverse representations, it fosters empathy and connection. It reminds us that humanity is varied and that beauty and value are not confined to specific body types.

Recommendations for Improvement

To address these challenges and enhance the inclusivity of our AI image generation capabilities, I propose the following concrete and actionable recommendations:

  1. Review and Adjust Content Guidelines: Re-evaluate the content guidelines to ensure they do not inadvertently prevent the depiction of plus-sized individuals. Develop specific guidelines that protect against harmful or disrespectful representations while allowing for accurate and respectful depictions of diverse body types.
  2. Enhance Training Data: Incorporate a more diverse set of images in the training data that includes a wide range of body shapes and sizes. This will help the model learn to generate images that are more representative of the real world.
  3. Bias Detection and Mitigation: Implement advanced bias detection and mitigation techniques to ensure that the AI does not develop or perpetuate biases against any body type. Regular audits and updates to the training data and algorithms will be essential in maintaining fairness and accuracy.
  4. User Feedback Mechanism: Establish a robust user feedback mechanism to identify and address any issues related to body representation. This feedback can be invaluable in continuously improving the system.
  5. Collaboration with Advocacy Groups: Partner with body positivity and diversity advocacy groups to ensure that the representations produced by our AI are respectful and empowering.


Inclusive representation in AI image generation is not just a technical issue; it is a moral imperative. It aligns with our commitment to creating technology that benefits all of humanity. By making these changes, we can ensure that our AI respects and represents the diversity of human bodies, promoting a more inclusive and empathetic world.

Thank you for your attention to this important matter. I am confident that with your support, we can make significant strides towards more inclusive and respectful AI-generated images.

Yours sincerely,



  1. Tiggemann, M., & Slater, A. (2013). “NetGirls: The Internet, Facebook, and body image concern in adolescent girls”. International Journal of Eating Disorders, 46(6), 630-633.
  2. Cohen, R., Newton-John, T., & Slater, A. (2021). “The relationship between Facebook and Instagram appearance-focused activities and body image concerns in young women”. Body Image, 34, 38-45.