AI’s gender bias issue is well-documented. But it’s part of a broader problem: systems reinforcing outdated stereotypes across gender, race, class and more. These old patterns persist and worsen, with studies showing they’re not just embedded in systems, but reinforced in audience behaviour.

You’d expect that with AI under constant global scrutiny, especially since going mainstream, we’d see progress. But OpenAI’s launch of Sora, and backlash over its portrayal of women in only domestic or passive roles, suggest otherwise.

So what’s going wrong? Are those driving AI simply unable or unwilling to fix its representation problem? If so, who will?

Part of the challenge is the dominant mindset, which sees technology as neutral - tools reflecting intent, not shaping it. This creates blind spots around how bias and cultural assumptions are embedded into systems that shape our every day worlds.

While technologists have struggled to course-correct, creatives could help steer AI onto a more inclusive path. Working at the intersection of culture, business, and technology, they understand how what we create reflects and continually shapes culture, through empathy and co-creation.

Female and marginalised communities' representation, and the transparency gap in AI

Unlike industries like engineering and healthcare, which follow long-established ethical guidelines, AI remains largely self-governed. Developers and business decision makers decide how systems are trained with little external accountability, meaning outcomes often reflect dominant perspectives of those building them.

This lack of transparency is especially damaging for gender representation and risks undoing decades of progress. When AI is fed biased content, from underrepresentation of women in professional roles to over-sexualised imagery, it simply mirrors those distortions. Without strict guardrails, it’s hard to trace or challenge outputs.

Too often, short production cycles and cost-to-value pressures limit design to ‘smoothing things over’, for dominant users, rather than questioning the systems themselves. But, business leaders have a valuable opportunity: to give designers space to influence AI’s structural fabric. That means supporting teams that meet practical and financial outcomes, while embedding cultural values like trust, inclusion and equal participation.

This is where design thinking can shift the conversation. Creatives, especially those trained in universal design, are skilled at building inclusion from the outset. By questioning omissions, challenging outputs, and demanding transparency, they can help steer AI toward more representative outcomes.

The younger generation, AI, and built-in ethical guardrails

AI is playing an increasingly prominent role in shaping culture. From voice assistants to AI-generated social content, younger generations grow up surrounded by systems influencing how they see themselves, others, and the world. With omniscient internet exposure, this influence starts earlier and runs deeper than ever.

Without intervention, AI will hard-code narrow biases, turning stereotyped views of gender, race, and identity into default settings. Combined with increasing misinformation, there’s a real danger of passing harmful narratives to future generations, disguised as neutral outputs.

This makes inclusive design a cultural problem as much as a technical one. If today’s tools shape tomorrow’s values, they must be built on foundations of diversity and fairness. This work starts with training: using representative datasets reflecting a broader range of identities and experiences. When you don’t feed AI harmful information, it is less likely to generate harmful results.

But datasets alone aren't enough. We also need people who can question the narratives AI systems are built on. And creatives are uniquely equipped to do just that. Skilled at decoding culture, they can bring the same critical lens to AI: asking what stories are being reinforced, and how we can design systems that support healthier, more equitable perceptions for future generations.

Ethical guardrails - what, how and why

Needing ‘ethical guardrails’ for AI isn’t new, but putting them into practice remains a challenge. As AI becomes more autonomous and shapes daily decisions, clearer mechanisms are essential to guide tools behaviour, who they serve, and priorities.

This responsibility goes beyond engineers. Bias audits, explainability, and inclusive teams are stronger when designers are involved. They uncover problematic patterns, test assumptions, and bring cultural literacy to a space still dominated by technical logic. Ongoing feedback loops prioritising underrepresented voices and clear accountability frameworks for failures are equally important.

Creative input already proves valuable: inclusive avatars in gaming, diverse illustrators shaping generative art, and copywriters refining AI-generated tone. These examples show the potential when creatives help shape AI, not just react to it.

A call to creatives

AI’s gender bias and lack of transparency aren’t inevitable; they’re challenges we can solve. Progress requires stepping beyond traditional tech silos and embracing diverse voices and skillsets.

Creative professionals are more than just storytellers. They are critical to building AI models that reflect the world that we want, not the one that we are heading towards.

By demanding transparency, designing for informed, inclusive user experiences rather than convenience for dominant audiences, pushing for inclusive datasets, and collaborating on ethical guardrails, creatives can shift AI from biased automation to equitable innovation.

The future of AI depends on better design - smarter algorithms follow. With creatives who can think critically at the table, technology can empower everyone, not just the privileged few.