AI and AGI are revolutionising the art scene in Southeast Asia, enabling artists to explore new mediums and reimagine their cultural heritage., AI-powered tools like Stable Diffusion and Midjourney are helping artists create photo-realistic images and videos, pushing the boundaries of creativity.,AI art generation raises questions about copyright, authorship, and originality, with legal uncertainties surrounding the use of these technologies.
The AI Art Revolution in Southeast Asia
Artificial Intelligence (AI) and Artificial General Intelligence (AGI) are transforming the artistic landscape in Southeast Asia. These technologies are providing artists with new mediums to express their ideas and reimagine their country's history. One such artist is Patrick Cabral, a multi-disciplinary artist from the Philippines who uses AI software like Stable Diffusion and Midjourney to create photo-realistic images of pre-colonial Philippines in a contemporary setting.
AI's Role in Reimagining Cultural Heritage
For Cabral, AI tools offer a unique opportunity to explore the question of how architecture and fashion would have evolved if the Philippines weren't colonised. However, he faced cultural challenges while using these programmes. AI algorithms often failed to understand regional concepts, resulting in inaccurate depictions of Southeast Asian clothes, people, or architecture. To overcome this, Cabral created lines of code to train his AI to recreate Filipino motifs and landscapes.
AI and the Mainstream Art Scene
AI's influence extends beyond individual artists. Galleries in the region are also capitalising on AI's increasing accessibility. For instance, the Singapore Art Museum launched 'Proof of Personhood,' an exhibition exploring the interplay between art, identity, and AI. This exhibition sparked conversations about the sentience of AI and its role in art among visitors of all ages.
The Dark Side of AI Art
Despite the benefits, concerns about copyright and human replacement have caused some to question if the AI race is accelerating too fast. The ease with which AI can replicate artists' styles poses a threat to originality in the market. In 2023, a group of USA-based artists launched the first class action lawsuit against several AI generators, including Midjourney, DeviantArt, and Stability AI, accusing the companies of copyright infringement. Warner Bros also took Midjourney to court more recently.
The Legal Grey Area
The question of who owns the copyright of AI-generated art remains a legally grey area. Ronald Wong, Deputy Managing Director at Singaporean law firm Covenant Chambers, notes that while there is a requirement for human authorship under the Singapore Copyright Act, who that human author is remains unclear. This could be intentional, as the review committee deliberately left open the question of computer-generated work and copyright ownership. For additional context on intellectual property rights in AI-generated works, a good resource is the World Intellectual Property Organization (WIPO) document on intellectual property and artificial intelligence.
AI in Art: A Double-Edged Sword
While AI and AGI are revolutionising the art scene in Southeast Asia, they also pose legal and ethical challenges. As these technologies continue to evolve, it is crucial to address these concerns and ensure that they are used responsibly and ethically.
Comment and Share
What do you think about the role of AI and AGI in art? Do you believe these technologies are a boon or a bane for the artistic community? Share your thoughts in the comments below and don't forget to Subscribe to our newsletter for updates on AI and AGI developments.
Latest Comments (4)
Cabral use fine-tune with his own code for regional context. This custom model approach very useful for production, especially for domain-specific tasks. I wonder, he share this fine-tune data? Or it stay private to him? Data sharing is big challenge for small language models in niche areas.
Patrick Cabral creating lines of code to train his AI for Filipino motifs" - this is exactly the kind of customisation we're seeing demand for in compliance automation. off-the-shelf models are never quite right for local nuances.
Elaine Ng (@elaineng) The point about AI algorithms failing local nuance, like with Patrick Cabral's work on pre-colonial Philippines, is precisely what we're seeing in broader digital media consumption patterns. It's not just about an algorithm failing, but about how these models reflect (or don't reflect) diverse cultural datasets. Definitely something I'll bring up in my next seminar on cultural representation in AI.
Interesting to see how Cabral customized Stable Diffusion for Filipino motifs. We've hit similar walls at Tokopedia trying to generate localized content with off-the-shelf AI models. It always needs extra training data or specific fine-tuning to really nail the Indonesian aesthetic, otherwise it looks… off. Makes me think about infrastructure for regional AI models.
Leave a Comment