If you’re interested for more, check out my course on this via this link: https://bit.ly/SB_AISEA101

Disclaimer

Note: knowing the rapid development of GenAI, things may have changed since the publication of this article. But as of the time of writing, which is August 2024, the content mentioned below is as true as it can be (with limitations to my own personal experiences).

Introduction

While I’ve used Generative AI extensively for data analysis, its potential goes far beyond just crunching numbers. I’ve been diving into how AI can transform different areas of the creative industry. Whether it’s designing visually striking posters, streamlining marketing campaigns, or boosting productivity by automating repetitive tasks, AI has become an incredibly versatile tool in my creative toolkit. But among these many applications, what really fascinates me is how these tools can be harnessed to create content that resonates deeply with audiences in Southeast Asia, and in my case, especially in Brunei and Malaysia.

But here’s the catch: as amazing as AI can be, it often comes with built-in biases that don’t always mesh with our local context. These aren’t just minor glitches—they can significantly impact how content is perceived. For example, an AI trained predominantly on Western data might not fully grasp the subtleties of our languages, traditions, or even aesthetic preferences. The result? Content that feels a bit off or, in some cases, completely misses the mark when it comes to cultural relevance – like trying to tell a local story through a foreign lens—it might look okay, but it doesn’t quite capture the essence.

Why Local Context Matters

Let me give you a couple of examples to bring this to life. Imagine using AI to generate a marketing campaign for a traditional festival in Brunei. The AI might pull from a global dataset and suggest visuals, slogans, or themes that are totally out of sync with the festival’s cultural significance. Or think about using AI to draft content in Malay. If the AI has been trained primarily on Western linguistic patterns, the text might be grammatically correct but still feel stilted or unnatural to a native speaker.

These aren’t just hypothetical scenarios—I’ve faced similar issues in my work. For example, when designing promotional materials for a local event, the AI-generated content initially leaned heavily on Western imagery and color schemes. It took a lot of fine-tuning and prompt engineering to get the AI to produce something that felt authentic and relevant to the local audience. This experience really drove home the point that while AI is powerful, it needs a guiding hand to ensure the content it generates truly resonates with the people it’s meant for.

The Hidden Bias in AI

In my experiments and various projects, I’ve spent a lot of time with several Generative AI tools, including ChatGPT-4o, Perplexity, Midjourney, Suno AI, Kaiber AI, DALL-E, Microsoft CoPilot, Google Gemini, Luma AI, Runway, LeiaPix (now immersity AI), and Microsoft Designer – most if not all with paid subscriptions. While these tools are incredibly powerful, they are largely trained on datasets that are predominantly Western. This influences the AI’s output in ways that don’t always hit the mark for Southeast Asian audiences.

Imagine trying to use these tools to capture the essence of our culture—like trying to frame something deeply familiar with a lens that wasn’t designed for it. The result might be technically correct but it often lacks the emotional and cultural depth that makes content truly connect with our audience. This gap isn’t just a minor inconvenience; it can be a significant barrier to creating content that feels authentic.

In response to these challenges, I’ve made it a point in my workshops to teach participants how to navigate and overcome these biases. The key lies in prompt engineering—essentially crafting specific, detailed instructions for the AI to follow. By refining how we interact with these tools, we can coax them into producing content that not only fits but also resonates deeply with our local culture.

For instance, when generating images using DALL-E or Kaiber AI, I encourage users to include cultural references, local symbols, or specific color palettes that are significant in Southeast Asia. This might mean telling the AI to use traditional motifs or ensuring that the visual style aligns with local aesthetics. Similarly, when writing text in Malay or another local language using ChatGPT or Google Gemini, it’s important to provide context that the AI might not inherently understand, like regional dialects or idiomatic expressions.

But prompt engineering isn’t just about being specific; it’s also about being creative. I’ve shown participants how to experiment with different approaches—like blending local references with universal themes—to create content that feels both familiar and fresh. And sometimes, it’s a process of iteration. The first few outputs from an AI tool might not be perfect, but by reviewing and refining the prompts, we can gradually shape the content into something that truly reflects our cultural identity.

Why Ethics Matter in AI Use

Another big part of this journey has been exploring the ethical side of AI. In Southeast Asia, where religion and culture are deeply intertwined, AI doesn’t always get these nuances right. For instance, I’ve encountered situations where AI-generated religious content, such as recitations of sacred texts, could easily be misunderstood or misrepresented, especially by younger audiences who might not have the knowledge to discern inaccuracies.

In one of my experiments, I asked the ChatGPT voice assistant to recite the Al-Fatihah in Arabic. While the technology was able to produce the recitation, it raised significant ethical concerns. The idea of an AI generating religious content poses questions about accuracy, authenticity, and the potential for misuse. If children, who might not fully understand the significance or correct interpretation of such recitations, are exposed to AI-generated religious content, there’s a risk of spreading misinformation or fostering a superficial understanding of deeply sacred practices.

This is why I always encourage open discussions in my sessions about the responsibility we have when using these tools. It’s crucial that we are aware of the potential for AI to inadvertently spread misinformation or cultural insensitivity, especially when dealing with content that is as sensitive and important as religion. Taking proactive steps to prevent these issues—whether through careful prompt engineering, setting strict guidelines, or advocating for better-trained AI models—is essential to ensure that AI is used responsibly and ethically in our region.

Bringing AI into the Creative Industry

On the flip side, AI has been a game-changer for me in the creative industry. I’ve used Generative AI for a wide range of creative projects, from composing music and writing lyrics to generating videos, templates, and graphics for books. These AI-generated elements have then been combined and refined using tools like Photoshop, Canva, Adobe Premiere, and CapCut, allowing me to produce polished, professional-quality content with remarkable efficiency.

However, even with all this power at my fingertips, I’ve learned that staying mindful of cultural context is crucial. For example, one project involved using DALL-E to generate an image of a man. I prompted the AI with specific local context, hoping to get a result that felt authentically Southeast Asian. The AI delivered a decent image, but there was a catch—it automatically added a beard, assuming that a Muslim man must have one. This led to a bit of a humorous moment where I found myself in Photoshop, carefully “shaving” the AI-generated beard off the man’s face and adding a traditional Malay “songkok” or “kopiah” to complete the look.

In another instance, I asked the AI to create an image of a Muslim woman wearing a headscarf. While the AI got the basics right, the woman’s features ended up looking very Western or Arabic, which didn’t align with the Southeast Asian aesthetic I was aiming for. So, I had to manually edit her eyes, lips, and other facial features in Photoshop to make her look more representative of our local culture. It’s moments like these that remind me how important it is to be aware of AI’s limitations and biases—and why ethical, responsible use of AI is so vital in our work.

Another key aspect of ethical AI use is transparency. When I use AI to generate content, I make it a point to label it as such. This isn’t just about giving credit where it’s due; it’s about maintaining trust with the audience. People deserve to know when content has been AI-generated so they can make informed decisions about how they engage with it.

Building a Community Around AI Awareness

One of the most exciting projects I’ve launched is my Instagram account, Shutter Speech (please search and follow!). This platform is dedicated to raising awareness about AI bias and promoting the responsible use of Generative AI. But it’s not just about pointing out the problems—it’s also a practical resource for creatives who want to produce content that’s truly relevant to Southeast Asia.

On Shutter Speech, I don’t just share the prompts I’ve used; I also walk my followers through the entire creative process, detailing the AI tools involved and the steps I took to refine the outputs. Whether it’s generating images, videos, or music, I aim to demystify the process and show how you can use AI in a way that respects and reflects our local cultures.

The goal of Shutter Speech is twofold. First, it’s about raising awareness of the biases inherent in many AI tools—biases that can subtly shape the content we create if we’re not careful. Second, it’s about empowering others by sharing my process, so they can learn from it and apply it to their own projects. By building a community around these ideas, I hope to foster a more informed and responsible approach to AI in our region, where creatives can collaborate, share, and innovate while staying true to our cultural roots.

The Importance of Transparency in AI-Generated Content

I mentioned transparency earlier and thought I’d also create a section specifically on it. Transparency is a cornerstone of ethical AI use, especially in the creative industry. As AI continues to play a larger role in content creation, it’s crucial that we are upfront about when and how these tools are used. This isn’t just a matter of ethics; it’s about building and maintaining trust with our audiences.

Generative AI is still relatively new in Southeast Asia, and in my workshops, training sessions, and professional work, I’ve noticed that there’s still a significant lack of awareness and understanding around the importance of labeling AI-generated content. I’ve heard AI-generated content on radio stations, seen it on Instagram, and even encountered AI-generated songs created by Bruneians—from individual creators to large, well-known companies—without any indication that AI was involved. This lack of transparency can be problematic, as it may mislead audiences or diminish the authenticity of the content.

When people engage with content, they have the right to know if it was created by a human or generated by AI. This transparency allows them to make informed judgments about the content they’re consuming. For instance, if an audience knows that a particular image, video, or piece of music was AI-generated, they might view it through a different lens, considering the technological influences behind it.

In my own work, whether it’s creating marketing materials, generating visuals, or composing music, I make it a point to clearly label AI-generated content. On my Instagram account, Shutter Speech, I don’t just showcase the end results; I also explain the role AI played in the creation process. This openness not only helps demystify AI for others but also encourages a culture of honesty and integrity in the creative community.

Transparency also fosters a deeper understanding of AI’s capabilities and limitations. By being clear about what was AI-generated, we help set realistic expectations and contribute to a more informed discussion about the technology. This is especially important in a region like Southeast Asia, where cultural and ethical considerations are deeply intertwined with content creation.

Ultimately, being transparent about the use of Generative AI is about respect—respect for our audience, for the creative process, and for the cultural values that we aim to represent. As AI continues to evolve, maintaining this transparency will be key to ensuring that these powerful tools are used responsibly and effectively in our creative endeavors.

Rindu Satay Madura link to song in Suno AI (see screenshot below).

Moving Forward Together

So, where do we go from here? We’re really just scratching the surface when it comes to understanding and adapting AI for Southeast Asia, and it’s something we’ll need to figure out together. As AI keeps evolving, it’s important that we keep asking questions, experimenting, and making sure the tools we use reflect the rich diversity of our region.

For those of us in the creative industry, this isn’t just about cranking out better content—it’s about creating work that really connects with people on a deeper level. By staying aware of AI’s biases and figuring out how to work with them, we can make our content more meaningful and impactful, truly capturing the essence of Southeast Asia.

I hope by sharing my experiences, I can spark some curiosity and a bit of caution as we all dive deeper into the world of AI. Let’s keep pushing the boundaries of what’s possible, but always with an eye on making sure what we create is as diverse and vibrant as the cultures we’re representing.