Can AI Create Literature? A Critical Look at Gen-AI and Indian Languages

• In English, the results can be startlingly coherent, but in Indian languages like Kannada, the output remains awkward and emotionless.
• The boundaries between human and machine authorship are blurring, but true literature still belongs to the human imagination.
• While AI augments literature, it cannot replace human creativity rooted in emotion, culture, and imagination.
• For India, the path forward requires strategic investment in native-language AI to preserve its civilizational richness and linguistic diversity.

With the rapid advancement of generative artificial intelligence (Gen-AI), a powerful technology capable of creating poems, novels, and essays within seconds, a fundamental question now stands before us: Can generative AI create literature just like a human? In search of an answer to this, I recently conducted a small experiment.

I had been wanting to write a story for a long time. Since I mostly write non-fiction, this was quite a difficult task for me. I had heard someone say that generative AI could write stories, poems, and novels. Instantly, I opened ChatGPT and asked it to write a short story in English. Apart from asking for a couple of details about the plot, it generated a beautiful story within just two minutes! When I first read it, I couldn’t help but feel it must’ve been written by a seasoned author. I then attempted the same thing in Kannada. The disappointing part was that the sentence construction by the AI in Kannada looked like the ‘feet of a hen’, awkward and unnatural. While it seemed to have mastered the art of natural language processing (NLP) in English, the same cannot be said for regional languages like Kannada, where it still seems to be a distant dream.

To sum it up, although AI appears capable of technically creating literature similar to humans, as of today, it remains more of a concept than a reality. Because literature is not just a collection of words, correct grammar, or well-formed sentences; it is the essence of human experience—a sea of emotions and a treasury of imagination.

The technology we have today can certainly be applied to many everyday tasks. For instance, in customer service centres, in drafting technical reports, or in writing high-quality emails, it fits in quite well. It reduces minor errors, saves time, and most importantly, improves efficiency in the systems we interact with. The days of struggling to find the right words are gone; the time we used to spend checking grammar and sentence construction is now saved. 

Today, AI can write letters, applications, reports, and messages instantly, tailored to the status, position, and personality of the person before us. All we need to do is provide direction; it takes care of the rest—grammar, language, and tone. However, AI doesn’t make any decisions on its own. Whether to accept or reject what it generates is entirely our responsibility. In my earlier experiment, the Gen-AI app once told me, “You are the boss of your story.” That’s all well and good in English. But when it comes to Kannada, in terms of word usage and sentence structure, AI is still in its early stages of learning the literary art.

A person named Oscar Schwartz once experimented on a hall full of people. He presented two poems—one written by a human and one by a computer—and asked the audience to guess which was which. At first, many could differentiate. But over time, it became increasingly difficult. The more literary content we feed into computers, the more sophisticated they become. If we were to feed the complete works of Shakespeare, George Bernard Shaw, or T.S. Eliot into AI, it could produce texts in their style and tone, complete with literary devices, capable of creating up to 10–20% of the emotional impact of their original works.

But it lacks genius. It cannot create something new. It cannot imagine new literary devices. The human brain—what a wonder it is! No one truly understands it. Even after centuries, we still don’t fully comprehend its structure or uniqueness. True, pure, and authentic literature is governed by the thoughts within the mind and illuminated by the light of the soul that shines upon those thoughts. Can such a profound, metaphysical creation ever be replicated by man-made artificial intelligence?

This is not merely pertinent to the Kannada language alone; it applies equally to all regional languages of our country—and indeed, of the entire world. Even though English may seem dominant to us, over 7,000 languages are still actively spoken across the globe today. Each of these languages possesses a unique and rich literary tradition. In such a context, it has become imperative for us to examine how creative artificial intelligence can impact these diverse linguistic worlds. Moreover, it is equally crucial to consider how we safeguard the language that forms the very root of our civilisation. Let us take this opportunity to reflect upon these pressing concerns.

Impact of Gen-AI on Indian Languages

Generative AI (Gen AI) is significantly impacting Indian languages by enhancing accessibility, preserving linguistic diversity, and creating economic opportunities, but it also poses challenges like potential erosion of regional languages and data scarcity. India’s linguistic landscape, with over 121 languages and 22 constitutionally recognised ones, presents unique opportunities and hurdles for Gen AI adoption. 

While Gen AI holds immense promise for Indian languages, its impact is uneven. Initiatives like Bhashini, AI4Bharat, and BharatGen are commendable for focusing on open-source models and inclusivity, but their scale and speed lag behind global counterparts. The reliance on English-based LLMs as a foundation (e.g., layering Indian language capabilities on Llama or ChatGPT) may limit cultural fidelity and perpetuate dependency on foreign technology.

Moreover, the economic narrative (e.g., $359-438 billion GDP boost) assumes widespread adoption and infrastructure growth, which may be optimistic given current funding and compute constraints. The “jugaad” mindset—India’s knack for innovative problem-solving—has driven creative Gen AI applications, but systemic challenges like data scarcity and regulatory gaps require long-term investment.

The risk of linguistic erosion is real but overstated in some narratives. Indian languages are deeply tied to cultural identity, and grassroots efforts (e.g., Karya’s data collection) suggest strong community interest in preserving them. However, without sustained government and private sector support, low-resource languages may struggle to keep pace with dominant ones like Hindi or Tamil.

Let’s discuss in detail the opportunities and threats to Indian regional languages because of this revolutionary Gen-AI.

Opportunities from Gen-AI 

1. Breaking Language Barriers and Enhancing Accessibility:

A. Multilingual AI Models: 

Gen AI is enabling the development of models tailored for Indian languages, such as Sarvam AI’s Sarvam 2B (supporting 10 Indian languages) and BharatGPT (12+ Indian languages for text and voice). These models facilitate translation, transliteration, summarisation, and voice-based interactions, making digital services accessible to non-English speakers, who constitute over 80% of India’s population.

B. Voice-Enabled Interfaces: 

Given that many Indians prefer oral communication (due to low literacy rates or oral language traditions), Gen AI-powered voice assistants like Sarvam Agents and Hanooman are revolutionising customer service, education, and government scheme access through platforms like WhatsApp and telephony. For instance, Sarvam’s voice AI can handle tasks like booking tickets at a cost as low as 1 rupee per minute.

C. Government Initiatives: 

The Bhashini program, backed by the Indian government, is creating open-source datasets for Indian languages to train AI models. Projects like Jugalbandi, a chatbot for government schemes, leverage Gen AI to provide information in multiple Indian languages, enhancing inclusivity.

2. Economic Opportunities:

A. Market Growth: 

Gen AI is unlocking significant economic potential in sectors like e-commerce, education, and media. For example, improved translation tools could boost e-commerce sales by 20%, tapping into a $4-10 billion market, while online education and media translation are projected to grow into multi-billion-dollar industries.

B. Content Creation: 

India’s 100 million content creators are using Gen AI tools (e.g., text-to-video platforms like Luma and Runway) to produce high-quality content in regional languages, driving the creator economy, projected to grow from $30 billion to $480 billion by 2035.

C. GDP Impact: 

An EY report estimates that Gen AI could add $359-438 billion to India’s GDP by 2030, with a cumulative impact of $1.2-1.5 trillion over seven years, driven by applications in services, education, and healthcare.

3. Preservation and Promotion of Linguistic Diversity:

A. AI4Bharat’s Contributions: 

The research lab at IIT Madras has developed open-source models like IndicBERT, IndicBART, and IndicTransv2, covering all 22 scheduled Indian languages. These models support tasks like translation, speech recognition, and text-to-speech, preserving languages with limited digital presence.

B. BharatGen Initiative: 

Launched in 2024, this state-funded project focuses on creating generative AI systems for Indian languages, emphasising data-efficient learning for low-resource languages. It aims to democratize AI through open-source foundational models.

C. Cultural Nuance: 

Unlike traditional translation tools that lose context by pivoting through English, native language models (e.g., Hanooman’s Hindi model) capture regional variations and code-switching (e.g., Hinglish), ensuring culturally relevant outputs.

4. Innovative Applications:

A. Education: 

Gen AI enables real-time lesson translation and personalised learning, bridging language gaps in rural areas.

B. Healthcare: AI-powered diagnostic tools and chatbots in regional languages improve access to medical services in remote areas.

C. Agriculture: Platforms like AgNext and CropIn use Gen AI to provide crop health insights in local languages, aiding farmers.

D. Legal and Governance: AI translation tools are used in courts (e.g., EkStep Foundation’s tools at the Supreme Court) and for delivering government services digitally.

5. Job Creation:

Gen AI is spurring new roles in data annotation, AI auditing, ethics, and language processing. Initiatives like Karya pay rural workers to generate speech data (e.g., $40 per hour for Odia speech data), boosting livelihoods.

Now let’s look into challenges and risks.

Challenges and Risks

1. Risk of Linguistic Erosion:

If robust AI models for Indian languages are not developed, users may increasingly rely on English-based models, threatening the relevance of regional languages. Vishnu Vardhan of SML Generative AI notes that without strong local models, English could dominate digital interactions over the next decade.

Global LLMs like ChatGPT and Google Bard are primarily trained on English, covering fewer than 100 of the world’s 7,000 languages, which marginalizes Indian languages with limited digital footprints.

2. Data Scarcity:

Many Indian languages have oral traditions, lacking extensive written records or digital datasets. Collecting data for less common languages is labour-intensive and costly. For instance, building datasets comparable to those for English-based LLMs could take a decade.

Code-mixing (e.g., mixing Hindi and English) and dialectal variations (e.g., five dialects in Muzaffarpur) complicate model training.

3. Ethical and Cultural Concerns:

Gen AI can “hallucinate,” producing inaccurate or culturally insensitive outputs, especially in translation or content generation for Indian languages. This risks spreading misinformation, particularly in sensitive domains like legal or medical translations.

Translating cultural nuances accurately remains a challenge, as pivoting through English often strips away context. Native language models are needed to address this, but they require significant investment.

4. Infrastructure and Funding Gaps:

Developing large-scale LLMs for Indian languages requires substantial computational resources and funding, which Indian startups often lack compared to global giants like OpenAI. Analysts note that India’s 1,500+ AI startups have not yet produced a globally competitive LLM.

High compute costs and limited investor conviction deter startups from building foundational models, pushing them toward applications built on existing global LLMs.

5. Regulatory and Ethical Gaps:

The absence of robust AI governance frameworks raises concerns about data privacy, algorithmic fairness, and transparency. India needs policies to ensure ethical AI development, especially for sensitive applications like legal or healthcare translations.

Gen-AI Is Reshaping Literature: A global perspective.

The literary world, long guided by the pen and imagination, is witnessing a quiet but profound revolution—one powered by generative artificial intelligence (Gen AI). From Tokyo to Toronto, and Cairo to Copenhagen, writers, educators, and publishers are exploring how machine intelligence can co-author, analyse, and preserve literature in ways never before imagined.

1. AI Joins the Writing Desk : 

Across continents, authors are welcoming AI into their creative processes. Tools like ChatGPT, Sudowrite, and Jasper are no longer novelties but assistants in drafting plotlines, suggesting dialogue, or even writing entire chapters. In Japan, one AI-assisted novel made waves in a literary contest, while European authors are increasingly blending machine-generated text into their work.

2. A New Chapter in Literary Analysis :

Beyond writing, AI is becoming a key player in literary scholarship. In universities worldwide, researchers use AI to analyse vast literary corpora, tracing thematic trends, authorial styles, and historical shifts. Educators are also adopting AI to teach literary devices, improve reading comprehension, and support student writing, especially in multilingual classrooms.

3. Breaking Language Barriers : 

Translation, once a barrier to global readership, is being reshaped by AI. Tools like DeepL offer more nuanced and culturally aware translations than ever before, bringing regional literature to international audiences. AI-generated audiobooks are now accessible in dozens of languages and accents, widening reach for both new and classic works.

4. Publishing in the AI Age : 

For both independent and traditional publishers, Gen AI is a game-changer. Authors use AI to draft blurbs, design book covers, and optimise marketing strategies. Some publishers are using AI to analyse reader trends and forecast potential bestsellers, blurring the lines between art and analytics.

5. Interactive and Experimental Frontiers:

AI is not only assisting literature; it’s reimagining it. Writers are experimenting with AI-generated poetry, chatbot fiction, and even choose-your-own-adventure narratives powered by real-time machine learning. In South Korea and Germany, interactive storytelling apps are thriving, merging literature with gaming.

6. Preserving the Past, Digitally: 

In the realm of preservation, AI is aiding in the digitisation and restoration of rare manuscripts and texts. Global institutions like the British Library and UNESCO are using AI to archive endangered languages, ensuring that the voices of the past are not lost to time.

As Gen AI continues to evolve, the literary world faces important questions about authorship, originality, and ethics. Yet, one thing is clear: the story of literature is being rewritten, and AI is holding the pen.

Latest Advancements in AI-powered literature

In the area of Gen-AI, everyday you see new light. Unprecedented technological growth has been seen in a year or two. Here let’s see some of recent developments.

1. Creative Collaborator – Mugafi 

AI tools are increasingly serving as co-authors, editors, and creative partners for writers.  Platforms like Mugafi enable authors to train AI models on their unique writing styles, ensuring consistency and authenticity in storytelling.  These tools assist with grammar, structure, and narrative flow, democratizing fiction writing and making it more accessible to a broader audience.  

2. AI-Powered Book Creation – Youbooks 

In nonfiction, platforms such as Youbooks are revolutionising content creation.  Youbooks allows users to transform their ideas into full-length books by integrating inputs from AI models like ChatGPT, Claude, Gemini, and Llama.  The platform ensures up-to-date facts through real-time web searches and captures the author’s unique voice by analysing writing samples.  Users retain full commercial rights to their finished books.  

3. Audiobook Narration

Audible has announced plans to introduce AI-generated voices for audiobook narration, offering over 100 AI-generated voices across multiple languages.  This initiative aims to broaden audiobook access globally and help creators reach wider audiences.  However, it has sparked backlash from authors, translators, and voice actors who argue that AI threatens the nuanced artistry of human narrators.  

4. Reviving Literary Legends

BBC Maestro has developed “AIgatha Christie,” a digital recreation of the famed crime novelist Agatha Christie.  This AI-powered project offers writing courses based on Christie’s texts and creative methods, providing insights into plot construction, character development, and storytelling techniques.  The initiative blends history, literature, and technology to preserve and share Christie’s legacy with new generations of writers and readers.  

5. Human-Like Communication

A recent study by City, St George’s, University of London, and the IT University of Copenhagen reveals that large language models (LLMs) can spontaneously develop human-like social conventions.  In a reward-based naming game, AI agents began to collectively adopt naming conventions through repeated interactions, mimicking human linguistic development.  This behaviour signifies a new frontier in AI safety and social integration. 

Conclusion

Creative artificial intelligence has emerged as both a boon and a potential challenge to the world of literature. While it offers remarkable possibilities, it also poses significant threats to regional languages and their literary traditions if not approached with timely awareness and care. The development and preservation of language is a vital responsibility of the state. Around the globe, many nations are actively working to raise awareness on this issue. From UNESCO to universities, numerous initiatives and programs are being undertaken by governments, organisations, and institutions.

If the growth of literature is stifled, it is tantamount to halting the progress of human civilisation itself. Therefore, it is imperative that we critically examine the various adverse impacts artificial intelligence may have on languages and literature, and formulate a farsighted plan to safeguard linguistic diversity and cultural heritage.

What we can conclude is that the creation of new, emotionally resonant, mesmerising literature that can captivate millions is not yet possible through artificial intelligence. Over time, our brains are becoming sharper. The cultivation of imagination must be integrated into our culture. Only then can we nurture and preserve literature—the most beloved and foundational element of human evolution—on this Earth forever.