Ask Gemini, Claude or Deep Seek to give you its views on the following:

Use the Prompt: “How does AI affect the perception of gender and sexuality”

Discuss  the possible implications of the answers provided examining how such information might affect your own perceptions of gender and sexuality?

You need to make at least one comment on posts by other students.

PLEASE REMEMBER TO TAKE SCREENSHOTS OF YOUR CONTRIBUTION TO THE DEBATE. THESE NEED TO BE PLACED AT THE END OF YOUR ASSIGNMENT

Although the majority of the response are rather generic regarding the expected influence AI could have, my eye caught the broad generalisation offered in the “Gendered AI Assistants” section. It mentions Siri and Alexa as its examples, but it fails to delve deeper into the topic and opts to offer information the general public would be able to link to their frame of reference. It failed to mention chatbots under its list of AI assistants. Since I was hesitant to go beyond the scope of the question and wanted to be sure that my understanding of Chatbots needing to be added to the scope of the reply, I offered a second prompt as follows:

(Snapshot B)

Within this list, it lists Joseph Weizenbaum’s seminal bot, Eliza (the bot I thought ought to be included in it’s first response) under the response of Real-World AI Assistants, nor does it mention Richard Wallace’s Artificial Linguistic Internet Computer Entity (Alice). This offers one of two conclusions:

1)      Deepseek failed to account for older information fitting to the topic, leading me to wonder how accurate and comprehensive the rest of the answers are, calling into question it’s ability to offer accurate information.

2)      It has a flawed view of what could be defines as an assistant, if Chatbots ought not to be included in the list at all.

(Snapshot C)

Furthermore, the heading being skewed to “Assistants” as generalisation, rather than accounting for gendered AI overall, offered a biased look at the search results in and of itself. AI is far more prevalent than purely in the pragmatic and applicable sphere or culture, so I would question it’s lack of reference to the pivotal role AI has played within fiction and narratives throughout the years. It makes no mention of the caregiver/saviour/provider roles Feminine or Female AI’s are portrayed in (From Star Trek’s ‘LCARS’, to Alien’s ‘Mother’) nor the generalised competent/assertive/antagonistic roles of Masculine or Male AI’s (2001: A Space Odyssey’s ‘HAL 9000’ or Marvel’s ‘Ultron’). Again, this calls into question whether the scope of the response to the prompt is fitting to the prompt itself, since there is no question that a portrayal of gender roles within fiction has an impact on societal gender norms, views and expectation.

When prompting DeepSeek with the given prompt, I made sure that I clicked on "DeepThink". DeepThink is a feature that allowed me to watch the AI’s thought process and see the logical steps it took to reach an answer. This was particularly insightful, as it not only allowed me to follow along with DeepSeek’s reasoning but also allowed me, in real time, to watch it backtrack on a thought and revise it. I did this because when comparing how AI and Human's have a certain perspective it is important to know what created that perspective and how it influences the answers given.

DeepSeek itself acknowledges that it contains bias in its training data as well as bias mitigation tools which have been inserted in post training by the developers in its answer. I saw both of these during the AI’s "thinking" process. Initially, the AI generated a thought based on the biased information it had been trained on, information that is largely sourced from the internet, where deeply ingrained prejudices like misogyny and queerphobia have been normalized and extensively used. However, I then observed it self-correct. This was likely due to the bias mitigation tools that are probably code implemented by its developers to detect a certain rhetoric and prompt a recalibrated thought that challenges the bias such as the mention of the "Taiwan square massacre" which gets a simple response of "Sorry, that's beyond my current scope. Let’s talk about something else".

This process of the AI self-correcting is fascinating because it mirrors what all critical thinkers are supposed to do. Regardless of intent or level of intellectual discipline, every critical thinker has inherent biases or preconceived notions shaped by exposure to societal norms, cultural narratives, and personal experiences. Just as AI internalizes biased data from the information it is fed, humans absorb biases from the media, family, friends, and all forms of communication. No form of communication, no matter how carefully created, is ever entirely free from bias. It is the responsibility of a critical thinker, much like the developers programming bias detection into AI, to recognize these biases and actively self-correct when forming thoughts and arguments. Over time, this process can become almost second nature, occurring subconsciously in the background of our reasoning.

A personal example of this would be my upbringing in a conservative family that held quite a rigid expectation of what it means to be a man or a woman. Initially, when I encountered men behaving in ways that contradicted these traditional norms, my 6th grade Natural Science teacher, I felt uncomfortable from an internal resistance born not from reasoned belief, but from ingrained bias of my upbring. My reaction was not based on any rational analysis of gender roles but rather on the ideas that had been imposed upon me by my family. However, through self-discovery, education and an expanded worldview, I came to understand how arbitrary these gender roles are. Today, they hold little significance for me, and I move through life in a completely unconventional masculine way. I no longer allow those biases to shape my decisions or perceptions.

That being said, overcoming personal bias does not mean ignoring its continued existence in others around us. Even though I have dismantled these biases in myself, I still recognize their presence in society and in all discourse. It remains my responsibility, as someone who has undergone this process of unlearning, to acknowledge bias when analyzing communication, decision-making, and power structures. An example of this is the systemic underrepresentation of women and queer people in positions of power. To engage critically with this reality, I must be cognizant that these biases exist and they might be based in abstract concepts but they have a direct tangible effect in shaping these power imbalances. By acknowledging the underrepresentation of women and queer people and understanding that is matches the ideals of different biases I can come to the conclusion that the exclusion of any marginalized group is not accidental is rather the result of long-standing biases that continue to dictate who is granted authority and whose voices are amplified.

In almost all ways, DeepSeek’s self-correction process exemplifies the kind of intellectual discipline required of us as critical thinkers. Just as AI must be programmed to detect and counteract bias, we must train ourselves to question, deconstruct, and refine our own thinking.

With this in mind, I opened Gemini in two separate tabs and gave it the same prompt in each. Not only did the phrasing of the responses differ across tabs, but certain central points were different too. The first response featured a heading entitled “Objectification and Sexualisation” while no mention of these issues appeared in the second. The second response did, however, feature an entirely unique section on “Privacy Concerns”. These differences raise concerns, like those mentioned by George, about the accuracy and comprehensiveness of any one response by an AI model. If I had only prompted Gemini once, I would not have encountered one of those sections. Who knows what I missed out on by neglecting to prompt a third time. The disparity in Gemini’s responses is illustrative of the nature of AI models. They are not thinking things with reason and opinions, but rather tools adept at condensing information. They will not necessarily answer the same question in the same way each time they are asked. I think that it’s important to recognise flaws like this as resulting from the imperfect nature of AI models as informative tools, and to subsequently think critically about anything they generate. 

In practice - to engage briefly with the content of Gemini’s answers rather than the meta-analysis - this means recognising instances of bias in AI output, and rejecting them. Such bias is obvious, for example, in content, whatever the form, that depicts women as stereotypically feminine, overly sexualised, or performing heteronormative gender roles. Bias could also arise in algorithms (media users may be exposed to gender-coded content), data analysis, or the absence of diverse representation. All of these things, among many others, could be incredibly harmful to people’s perception of gender and sexuality, and could perpetuate harmful ideologies such as sexism, homophobia, and transphobia. A resurgence of outdated, bigoted beliefs such as these would have dire consequences for the progress we have made towards a more equitable, accepting society. On the other hand, I do believe that there is hope. As Gemini points out, when AI models are programmed and used thoughtfully, they can create content that challenges stereotypes, impacting the perception of gender and sexuality in a positive way. AI can be used - as I would like to use it - to deconstruct archetypes and formulate new, innovative ways of representing gender and sexuality as changing and fluid. 

To return to the meta-analysis: I was struck by the irony of Gemini’s partial self-awareness. Gemini knew to label the reinforcement of stereotypes as belonging to “Risks and Challenges” rather than “Potential Benefits”; it acknowledges that inclusion is desirable and that biases are not, and yet this theoretical “awareness” is often insufficient to prevent it from perpetuating biases. My use of scare quotes illustrates the point I’m trying to make: that AI is not truly aware of anything. It has programmed parameters, but these are not yet finely-tuned (hence the “woke” images resulting from prompt transformation) and they are themselves subject to bias: as mentioned in Gemini’s response, AI development teams often lack diversity. Beyond these parameters, AI is trained on datasets that include both literature about the undesirability of bias, and literature that is itself biased. This results in Gemini’s ability to recognise that it ought not to reinforce bias, and then to reinforce it anyway. I find this fascinating and scary, and I think it speaks, again, to how vital it is to think critically about AI’s output. 

Considering all of this, I approach AI representations of gender and sexuality with caution. As someone who has spent considerable time thinking about my own gender and sexuality, I would like to think that my perceptions of them are nuanced and compassionate. To know that AI is subject to bias is to know to treat it critically, but it is also to recognise that there is an extensive, problematic history surrounding the perception of gender and sexuality which needs to be addressed. Reading Gemini’s answers, I was struck by the struggle going on behind the scenes. I was reminded (although I need little reminding) that gender and sexuality are contentious subjects about which so much harmful and ignorant content has been created. I was also reminded, though, of how far we have come in our perceptions of them and of how hard we are willing to work to not go back. AI models are not only the data they are trained on. They are also the interventions (which Aidan describes so well) that smart, thoughtful people are coding in order to write back against the data, to correct the mistakes we left there. I find the idea of these interventions quite poetic: attempts to wade back through the sins of the past - it’s all there, in the dataset, everything we’ve done wrong - and to say, in code: this is no longer the way we want to be; we can be better.

DeepSeek highlights the potential for AI to reinforce traditional gender roles, pointing out how biases in language and professions can be perpetuated by AI systems. Yet, it also acknowledges AI's capacity to foster inclusivity by training on diverse datasets, which could broaden our understanding of gender identities and sexual orientations. In contrast, Gemini stresses the risks of AI amplifying existing societal biases, emphasising that if AI models are trained on biased data, they may inadvertently perpetuate harmful stereotypes. What is particularly notable is the self-awareness demonstrated by both AI systems, as they recognise their own susceptibility to bias. This underscores the ongoing challenge of ensuring AI models don't merely reflect human prejudices, but rather contribute to a more equitable and inclusive future.

The implications of these AI perspectives on gender and sexuality have the potential to shape not only my personal views but also societal norms at large. While I strive to challenge stereotypes, I must acknowledge that AI can subtly reinforce preconceived notions. For example, when interacting with AI-driven content that frequently associates specific genders with particular roles, I might unconsciously internalise these portrayals as normative, even if they contradict my conscious beliefs. Over time, such subtle biases may influence my perception of gender and sexuality, potentially narrowing my understanding and reinforcing outdated stereotypes. This realisation underscores the importance of engaging critically with AI-generated content to prevent the unintentional adoption of biased views.

AI-driven algorithms, such as those used by platforms like TikTok, shape not just what I see but also what the broader public encounters. These algorithms recommend content based on user preferences, creating personalised feeds that cater to individual beliefs and interests. As a result, people are often exposed to media that reflects their existing views, reinforcing those perspectives on gender and sexuality. This creates a feedback loop, wherein users become increasingly immersed in content that aligns with their beliefs, further entrenching their views and making it more difficult to engage with differing perspectives. The algorithmic biases at play aren't limited to TikTok—any platform that personalises content, whether it be books, news, or movies, can limit exposure to diverse perspectives, reinforcing societal silos and narrowing our collective worldview.

AI's influence extends beyond social media and into creative fields like art, poetry, and literature, where human emotional expression plays a crucial role. While AI-generated content can be useful in some contexts, it can also be biased, just like the data it's trained on. AI often lacks the depth, emotional resonance, and nuance inherent in human-created works. Poetry and literature, for instance, are deeply rooted in personal experience, emotions, and perspectives. When AI generates these forms of art, it risks reducing them to mere patterns, influenced by existing biases, and devoid of the emotional connection that makes them powerful. This can lead to the reinforcement of gender stereotypes, such as portraying women in nurturing roles and men as leaders, racial and ethnic biases, where certain groups are linked to specific occupations or behaviors, socioeconomic biases, showing wealthier individuals as more competent or deserving, and cultural clichés, reducing complex cultures to oversimplified traits. This could lead to a creative landscape that feels more formulaic and less authentic, potentially amplifying stereotypical views and diminishing the richness of human artistic expression.

As AI continues to evolve, I find myself asking – is AI a tool for breaking down societal biases, or will it only mirror and magnify the prejudices already embedded in our systems?

Despite the negative implications, AI can promote inclusivity and fight against stereotypes. For instance, AI now discusses non-binary identities, which promote diversity and inclusivity. Reading outputs generated by Gemini and Claude makes one conclude that AI can influence society’s opinions on gender and sexuality. Spending time analyzing the inputs has made me aware that AI can mold biases, which raises a challenge for me not to take any input I come across. Instead, I should apply critical analysis skills, ensuring the selection of outputs that promote inclusivity and diversity

One could argue that it is just as much the answers Deep Seek will provide us when offering it this prompt as it is the ‘thinking’ behind the AI screen that matters. It is clear that Deep Seek, like humans, must weigh many aspects of its way of thinking to provide us with answers. It acknowledges that it can indeed be biased, but also has the scope to try to avoid gender biases. However, this again reflects on its maker and their spectrum of inclusivity, as well as what they view as biased and what they do not. An example of the ‘thinking’ Deep Seek considers is an all-male team of programme developers or development teams, which Deep Seek uses as an example of the chance of bias being programmed into the AI system due to a lack of diverse programming teams. It is interesting that Deep Seek recognises or is programmed to acknowledge that many phone assistants are designed to have female voices. Now, one is left to wonder whether Deep Seek was programmed to identify and recognise such biases within technology itself. However, since it is, in fact, the product of humans, it has its shortcomings, just as the human who created it does.  This debate made me think of the children’s series The Jetsons, which predicted many technological developments before they were produced or as they were being designed. The maid in the series is a female robot called: Mechano Maid 2000 – again, a female assistant, maid and nanny. Here we have this fantastic idea of technological development, but again, it has its maker's biases and shortcomings. I think, based on the answers Deep Seek has provided, it can help individuals become more inclusive; however, this all depends on the programming and the programmers. It is a human base from which this programme was developed, thus it is possible that which continues development this product could be a guide as to aid more closed-minded induvial view the world more inclusive, but again this requires the programmer to be an open-minded, inclusive and continuously improving developer with its product, as to keep up with the ever-changing world and our roles as humans in it. In conclusion, AI is the product of its maker, and if one is to question the AI, one must question its maker. AI has the potential to profoundly impact the perception of gender, whether for good or ill, which is ultimately up to the human being using it and the human being creating it.

AI generated content such as images or texts, can perpetuate harmful stereotypes if the underlying models are biased. This can influence how people perceive different genders and sexualities, especially if they lack real-world exposure or rely heavily on AI generated information. AI models have been shown to generate more diverse and higher status jobs for men in stories compared to women. This implies that men are superior to women. AI can create harmful or untrue ideas, stories or images that show gay people or people from certain ethnic groups in a bad or unfair way. This could include spreading lies or treating them poorly based on the identity.

Claude also has a similar view on the perception of gender and sexuality. Just like Gemini, Claude states that AI systems often reflect and sometimes amplify existing biases in their training data. When trained on texts and images from the internet, they can perpetuate stereotypical representations of gender and sexuality. For example, image generators might associate certain professions with specific genders or produce stereotypical depictions on LGBTQ+ individuals. Many AI personas are designed with gendered characteristics. The predominance of female-coded assistants (Alexa,Siri) may reinforce notions about service roles being feminine, though companies have been diversifying voice options in response to criticisms. AI systems moderate content across platforms, affecting which expressions of gender and sexuality remain visible.

Gemini explains the impact of AI in the perceptions of gender and sexuality as difficult and multi-dimensional, advancing both potential benefits and significant risks.

AI reinforces existing societal biases and this can result in AI forcing harmful stereotypes about gender and sexuality, like AI might associate certain professions and behaviour to a particular gender. This may pose danger to a certain extent, a facial or physique recognition can lead to misidentification or discrimination against individuals who do not conform to traditional gender norms.

AI lacks diversity in data and development. The underrepresentation of diverse gender identies and sexual orientations in AI can lead to systems which lack sensitivity or are discriminatory towards the affected groups, for example, non binary and LGBTQ. The results may lead to discriminatory outputs that provide biased recommendations or generate stereotypical contents.

AI often lacks diverse representation among its creators. This may lead to potential bias and failure to consider the needs of diverse group. The concepts of gender and sexuality are very diverse are culturally dependent. Using Gemini as a prompt, the influence of AI on gender and sexuality is a double edged sword which may sometimes be used to advance positive changes and also exacerbating current inequalities. Moreover, to strive for inclusivity and equity.

Ethically, AI systems lack privacy concerns. AI needs to ensure that its systems are designed to safeguard individuals from harassment and discrimination. Do away with harmful stereotypes or contribution to social inequality.

I believe that all the answers provided may to an extent be true, however, one needs read critically and not take all information provided as cut and dried. We should not lose sight to the fact that gender and and sexuality is fluid in its nature and that we should always ask questions such as, "Why? Why not? Yes, How? Raising awareness about potential biases by Policy Developers of AI and the public at large.

However, as an open-minded scholar, reading the Gemini’s answer in the question, it seems to me that there are biases when it comes to gender roles, and statistics play a big role on this information, it is not shocking to me that these biases are present. I think that biases is built into seemingly neutral machine learning systems, like AI. Because they are fed data to learn and predict outcomes regardless if it is harmful or misleading, which I believe should be addressed since most of the people who use AI thinks that the answers that are given to them are accurate. On the other hand, Deepseek discusses more than just data and statistics. It was mentioned that AI is also used as a tool for advocacy and inclusivity especially for discrimination of LGBTQ+ community in terms of having jobs and availing healthcare and the positive side of it is the effect of this advocacy and inclusivity programs. Aside from that, Deepseek also mentioned the risks such as deepfakes and fetishization that sadly it is also mentioned that it disproportionately targets women, LGBTQ+ and children, which I believe that it was also from data fed to the system.

After reading the answers on the AI platforms, there will be no changes when it comes to how I perceive gender roles. The information presented to me shows that AI are biased because they are human creations and the algorithm is fed data using historical discrimination and this results in a feedback loop whereby new AI systems are replicating old patterns, but, as a person of faith, I reflect and only take what will resonate on my beliefs, and of course, if you are specific in what you prompt onto AI, it will generate what you are looking for.

-Firdows

HOW DOES AI AFFECT THE PERCEPTION OF GENDER AND SEXUALITY

Artificial Intelligence like Claude AI is trained on extensive datasets of human-produced content, meaning we frequently reflect prevailing societal patterns, biases, and norms about gender and sexuality. However, AI may accidentally build up stereotypes or conventional perspectives.

Various AI systems are allocated gendered names, voices, or identities like virtual assistants with feminine voices, potentially stiffening association of service roles and particular genders. Moreover, in various customs, "Claude" is regarded as masculine, even though I do relate it with any gender because at the end of the it is an AI.

The manner in which AI systems execute content pertaining gender and sexuality portrays evaluation about which expressions are suitable, perhaps affecting what individuals   regard as normal or acceptable.

AI failed to develop guarded environments for individuals to investigate inquiries regarding gender and sexuality without fear of judgment, perhaps supporting self-exploration.

I think AI systems can offer insights into various gender identities and sexual orientations, which may amplify comprehension and a sense of embracing diversity.

The lack of representation of specific groups in training data can result in AI systems that are more adept at understanding and reacting to the expressions of gender and sexuality from dominant groups while sidelining others.

The connection between AI and human views on gender and sexuality is intricate and continually evolves. As AI increasingly merges with everyday life, its impacts will probably grow in importance and require continuous ethical evaluation.

AI’s perception of gender and sexuality come in both negative and positive effects; and these effects carry an intense implication for both society and the person using it. Deepseek does agree it reinforces stereotypes, such as depicting nurses as female and engineers as male. By doing this, it does not just reflect on these existing biases but systematizes them in a way that shapes everyday interactions. When hiring algorithms, chatbots or image generators default to gendered assumptions, they indirectly teach users that these patterns are unavoidable, or rather natural. The risk that comes with is that it creates a feedback loop, where AI affirms societal prejudices, making it difficult to destroy them. For some, especially for those still struggling to form their identities, constant exposure to these “norms” could make them internalize limiting ideas about gender roles or sexuality. Take for example, a non-binary person who constantly get misgendered by voice assistants, they might feel excluded; while a girl who constantly see AI-generated “scientists” as primarily male, might unconsciously look away from STEM fields. From all this, we can see that AI does not just describe reality, but it actively participates in constructing it.

On the contrary, AI also does have the potential to challenge the social norms to offer equality. AI can make use of tools like LGBTQ+ or non-binary voice assistants to validate identities of the marginalized. This will provide a representation that the traditional media often excludes. For those seeking self-understanding, particularly queer and trans individuals; virtual spaces where they can experiment with pronouns avatars can be transformative for them, as they carry or rather allows for low-stake identity exploration that might confuse them. However, this raises ethical questions about agency and authenticity. If an AI chatbot “affirms” a user’s gender identity, does that democratize access to support or replace human connection? Does it advance inclusivity or commodify identity for profit when it generates diverse narratives or imagery? The tension between AI as a tool for equality and exploitation depends on who is in control of these technologies, and on whether their design prioritizes equity over engagement or efficiency.

On a personal level, I think engaging in these realities forces us to evaluate how much of our own perceptions are shaped by algorithmic bias. As I was busy discussing the answers given by AI, I asked myself a question of whether my assumptions about gender roles stem from lived experiences or from AI-curated content like social media feeds that reinforces binary aesthetics, or voice assistance that are set to default as submissive, and feminized tones; and my answer was that they stem from lived experiences. For people questioning their sexuality, AI companions may provide a low-risk space for self-discovery, but they can also create unrealistic ideas about intimacy, since programmed responses can't match the depth of real human connection. The cultural consequences are undeniable: the more we integrate AI into daily life, the more its ingrained biases shape our perceptions unless we consciously challenge them. This requires critical thinking for choosing inclusive platforms, pushing for openness about how AI is trained, and staying aware of how technology often reduces the nuances of identity. In the end, AI's influence on gender and sexuality isn't predetermined; it's shaped by whether we use it to reinforce old norms or build a more flexible and just understanding of who we can be.

A study on the Gemini 2.0 Flash Experimental indicates that it assists in mitigating certain gender biases. Nonetheless, it adopts a more lenient stance on sexual content and permits more violent prompts. This prompts significant discussions regarding the normalization of harmful content and highlights the necessity for ethical standards in AI systems. Gemini aims to foster a conversational, inclusive, and user-sensitive environment. It underlines responsible and inclusive perspectives on gender and sexuality. By acknowledging nonbinary identities and LGBTQ+ experiences, Gemini contributes to creating a supportive atmosphere where users feel validated and empowered to explore nontraditional gender or sexual identities. On the other hand, DeepSeek presents a contrasting approach by emphasizing research-based or knowledge-driven AI. This strategy relies on data analysis and policy-oriented language to provide users with factual trends and cultural insights. While this methodology may seem more neutral or cautious, it encourages users to critically assess how gender and sexuality vary across different contexts and cultures. Gemini utilizes affirming language that emphasizes lived experiences, queer theory, and fluidity of identity. This approach helps normalize evolving identities and fosters open-mindedness and acceptance. Conversely, DeepSeek adopts a neutral tone based on statistics and legal or political matters, promoting an academic comprehension of gender and sexuality. I have noticed that this invites users to engage deeply with the subject. Culturally, Gemini benefits from a globally diverse dataset shaped by Western perspectives on gender diversity and sexual orientation. In comparison, DeepSeek taps into a range of geopolitical norms. This diversity can enhance discussions on gender and sexuality by offering wider viewpoints. User engagement also varies between these two models. When users engage with Gemini out of curiosity, it promotes exploration and affirmation. In contrast, DeepSeek stimulates critical thinking and provides sociological frameworks, helping users comprehend that gender and sexuality are shaped by external influences. This distinction allows Gemini to influence perceptions on emotional and social planes, while DeepSeek enhances intellectual and structural comprehension, enriching the user's learning experience. The impact of AI on perceptions of gender and sexuality presents both prospects and hurdles, making it vital to approach this constructively. When AI produces content that reinforces conventional gender roles or stereotypes, it provides an opportunity for us to reflect on our own biases and ensure that our perspectives remain open and inclusive. Engaging critically with AI-generated content encourages us to question underlying assumptions and motivates us to seek diverse viewpoints. I beleieve that being conscious of our language is also crucial. AI can serve as a valuable resource for fostering empathy and understanding by aiding us in learning about different identities and challenging our biases.

AI's impact on the perception of gender and sexuality is multifaceted and complex, presenting both opportunities and challenges. Gemini provided a breakdown of different perspectives on how AI may affect how gender and sexuality is perceived. Gemini reveals that AI reinforces and amplify existing biases and that result from biased training data. AI systems learn from the data they are trained on, if that data reflects societal biases regarding gender and sexuality, AI is likely to reinforce and perpetuate these biases. AI looks into tons of already existing information and if most of that information shows man in certain jobs and women in others, AI is likely to pick up that pattern. For example, if you ask AI to write a story about an Engineer it is likely for it to picture a male in the storytelling because many of the stories or old stereotypes shadow male Engineers. As much as AI reinforce and perpetuate existing biases, stereotypes and normative conceptions it can also be used to challenge and disrupt traditional perceptions of gender and sexuality. Gemini reveals that AI can be used to identify subtle biases that might be difficult for human to recognize. Its ability to analyze large datasets can help in understanding the nature of biases in language, media and other forms of data set. If AI could be exposed to a wide range of information including all kinds of people and relationships it can start showing a more realistic and diverse world instead of focusing on old stereotypes. AI generated content can be positive or negative based on the data AI has been exposed to, AI does not necessarily know what it means to be a man or women or anything in between it just follow a pattern. If the pattern is biased Things could get worse such as misidentification of people and assumptions based on limited information. This mainly affect transgendered people or those who do not fit in simple categories. How AI affect how we see gender and sexuality is a mixed bag. It can hold on to old ideas and stereotypes making them seem real but can also open our eyes to a more inclusive view.

At first glance this looks like an accurate and harmless summation of AI's role in shaping people's views but I suspect that this understates the influence of AI on our thinking.

If we don't question AI content, we will come up short as there are existing stereotypes/biases that reflect the views of the developers and policy makers; not just the users. An example is the popular virtual assistants - the default assistants are female unless intentionally overridden by the user. Furthermore, key professions are often stereotypically portrayed - doctors and nurses are often as portrayed as male and female, respectively and although there are AI tools that are designed to disrupt these stereotypes (e.g. Q), they are not the mainstream tools.

Beyond these stereotypes, I wonder how effectively and accurately gender representations within Southern Africa are presented. According to Scrums.com, (https://www.scrums.com/top-100-ai-companies-in-africa) the developers of AI are largely northern hemisphere based (US, China and UK are leaders in both patents and investment) while the Southern hemisphere AI landscape is "nascent".

If AI developers are predominantly CIS gender, heterosexual and Western, is it accurate to assume that AI is inclusive and unbiased?

According to deepseek, AI plays an important role in shaping how we think about gender and sexuality. It reflects what society believes, but also influences these beliefs. On one hand, AI can reinforce stereotypes if it's trained on biased data. This means it might promote traditional gender roles or assume everyone is straight.

However, AI can also be a positive force. When designed with care, it can help people explore their gender and sexual identities. For example, chatbots and virtual influencers can provide support and community. Unfortunately, AI also raises concerns, such as deepfake pornography and dating algorithms that can be unfair.

For LGBTQ+ individuals, AI can offer mental health support, but it can also censor important content. Moreover, I've personally seen instances where AI provides incorrect information, which can lead to the spread of misinformation, reinforce harmful stereotypes, and cause confusion or self-doubt. Given these limitations, we need to be cautious when relying on AI for personal and academic purposes. To make AI a force for good, we need to develop it responsibly, using diverse data, reducing bias, and regulating its use. By doing so, we can harness AI's potential to drive positive change and avoid harming people.

An example of such artistic work can be seen in the creations of artist J. The body of work produced by Rosenbaum demonstrates the interaction between computer vision technologies and concepts related to gender presentation as well as personal identity.

Through my engagement with visual art produced or classified by AI technologies, I have become increasingly conscious that these artworks often fail to encapsulate the emotional richness inherent in human gender identity experiences.

The ongoing evolution of artificial intelligence in its language processing and image recognition abilities demands that we as consumers maintain a watchful awareness regarding the ways these technologies influence our perceptions of gender and sexuality within artistic contexts. Through these actions we enable the development of an inclusive creative dialogue which acknowledges diverse identities while simultaneously recognizing the inherent limitations present in algorithmic perspectives.

The perception of gender and sexuality by AI can significantly influence how novels and poetry are interpreted, especially since literature often explores complex human identities. AI models are trained on existing human data, which can carry cultural or historical biases. If the data reflects limited or stereotypical views of gender and sexuality, AI might misinterpret or oversimplify nuanced representations in literature. Whe it comes to literature, it often includes fluid or non-traditional expressions of gender and sexuality. An AI with a narrow perception may fail to fully capture the emotional or symbolic depth of such characters or themes. AI used for translating or adapting literary texts might unintentionally erase or alter gendered or queer identities due to limitations in language models or translation rules.

The process of literary translation or rewriting through AI technology has the potential to dismiss or change gendered or queer identities because its language models and translation rules have certain limitations.

Reader Perception Becomes Altered when they use AI summary and analysis services since AI-based interpretation shapes how they understand literature in ways that both affirm and confront their current beliefs.

AI and the perception of Gender and Sexuality

AI systems can perpetuate existing biases present in training data. This can lead to misrepresentation or stereotyping of gender and sexual identities. For instance, AI models trained on biased datasets can reinforce stereotypes, such as associating certain professions with specific genders, which can negatively impact job opportunities and perpetuate existing stereotypes.

The lack of diverse representation in AI development teams can result in products that do not adequately reflect the experiences of all genders and sexualities. This is compounded by the fact that technology often shapes culture and its meanings, and without diverse input, AI can fail to challenge existing biases and instead reinforce them.

I think AI technologies, such as chatbots and virtual assistants, can influence how individuals perceive their own gender and sexuality through interactions and responses. Virtual assistants with female voices, for example, can emphasize traditional gender roles and contribute to the dehumanization and objectification of women.

AI should enhance, not undermine, individuals' autonomy in expressing their gender and sexuality. Ensuring that AI systems are designed to support rather than restrict personal expression is vital for maintaining individual agency.

AI can provide support for individuals exploring their gender and sexual identities through resources and community-building platforms. These systems can offer valuable assistance and create safe spaces for exploration and self-discovery.

The debate surrounding AI and the perception of gender and sexuality is multifaceted, involving ethical, social, and technological dimensions. It is crucial to approach this topic with sensitivity and a commitment to inclusivity, ensuring that AI serves to empower rather than marginalize.

As AI continues to evolve, the choices we make today will shape its future impact on society

Artificial intelligence represents cultural discourses surrounding gender and sexuality .In addition being a technological instrument. Traditional assumptions of nurturing, obedience, and service are reinforced by the feminine characteristics frequently incorporated into AI systems, such as voice assistants like Siri and Alexa. This design decision reflects patriarchal traditions and is based on Judith Butler's concept of gender as a performance.

In the meantime, popular media presents AI as sensual and hypersexualized in films such as Ex Machina and Her, where artificial entities are transformed into representations of human desire, posing difficult moral dilemmas about intimacy and permission. Not all representations, nevertheless, are constrained by these binaries. Not all representations, nevertheless, are constrained by these dichotomies. In keeping with queer and posthuman theories like Donna (1991) cyborg manifesto, which promotes erasing the distinction between humans and machines and upending traditional gender roles, efforts toward gender-neutral or non-binary AI challenge established categories . It shows that while AI presents a chance to reconsider and question conventional ideas of gender and sexuality, it also reflects and may even reinforce existing societal biases.

The goal of some technologists is to develop androgynous or non-gendered AI. Many users still view things via gendered glasses, though. This implies that human perception rather than technical design may be more important.

If I were to use AI to assist me in analysing and interpreting novels or poetry it would not give me the emotional depth I require from reading and analysing those readings myself, from my own subjective experience. AI often comes up with shallow or even inaccurate readings, particularly of works that challenge understandings of gender identity. AI can often misinterpret poetry readings to a large extent, whereas as a human, I am able to provide a critical, in-depth reading of the poetry, analyse every line, every word, and every single meaning as every word carries layered meaning, something AI might overlook, the same goes for artwork. I think that AI’s perception of gender and sexuality could take real, raw meaning out of novels, poetry and artwork. It can only analyse something to an extent, whereas as humans, we take it many steps further. I argue against using AI and relying on it for academic writing, although it is helpful to map out the structure of an essay, it completely flattens my own voice, I am quite descriptive when I write, I like to break down each sentence and reveal it’s layered meaning (this is also a downfall, sometimes texts are just texts and there is no hidden meaning) and AI is unable to capture this rich, layered analysis. Also, using AI models can create a false sense of understanding without actual learning involved. I believe it should not replace the reflective work in academia that ultimately defines meaningful engagement with all varieties of literature works.

AI's perception of gender and sexuality has no bearing on the literature I read or the art I view. I don't knowingly read or view creative works that have been generated by AI systems. If I did, I fear my perception of gender and sexuality would certainly be an outdated one - perhaps even harmful. Especially as the who behind AI affects the what. People are biased. It's our nature. It's not a leap to assume that developers are training AI systems in their unconscious biases. In a UN Women article it was noted that when women use AI to diagnose illnesses they often received inaccurate information because the AI was not written to include the female experience. Another point noted how an artist prompted AI to generate a story about a doctor and a nurse for their research and the results made the doctor male and the nurse female.

I have used AI to format my references. But the results still required minor punctuation adjustments. My intention is only to use AI in an administrative sense and not to generate creative answers.

{If I may share an anecdote.}

At my previous school, my department head and fellow colleagues all urged me to use AI to complete the unending admin-related tasks we kept receiving. This soon included lesson plans and presentations. It felt highly unethical, especially as the school positioned itself as THE school in the city. I kept thinking that I'm dulling my senses/skills, and cheating my students and their parents who had invested a considerable amount of money to send their children to the school. But it ended up being a choice between continue sleeping 3-4 hours a night or accept the times that I am living in. I still look back on that period with guilt.

Perhaps AI development should consider not only the gender and sexuality identity of all their users but also the social (cultural) systems within which AI is used?