"Digital inclusion in a multilingual society needs more than translation." Critically analyse.

India's Digital India BHASHINI platform now covers 36 Indian text languages and 23 voice languages [1]. The gap between those two numbers makes the point. Translation turns text into more languages. Real inclusion also needs speech, data, deployment and access, especially for minority languages.

Translation: necessary, and already showing results

• Core services: BHASHINI offers translation, speech recognition, speech synthesis and transliteration for public-facing services [1].
• Institutional use: Under the Sansad Bhashini MoU, Lok Sabha and MeitY are building tools from Parliament's own records [2]. This makes legislative content available in many languages.

Why translation alone falls short

• Voice for low-literacy users: Many citizens cannot read translated text. DIBD and IISc therefore built text-to-speech models for Sourashtra, a minority Indo-Aryan language spoken mainly in Tamil Nadu [1].
• Two-way interaction: The release covers TTS only, with no speech recognition for Sourashtra [1]. Users can listen to a service but cannot yet speak to it.
• Data scarcity: Large languages have existing text, such as parliamentary debates [2]. Sourashtra, with about 2.48 lakh speakers, needed a corpus recorded from scratch, and it has only 28+ hours [1].
• Representativeness: That corpus comes from just two voice artists and uses scripted or literary sentences [1]. The output may not sound like everyday speech across ages and dialects.
• Script plurality: Separate Tamil-script and Sourashtra-script models are needed [1]. This splits an already small user base.
• Model is not a service: The release names no department or scheme that will use the models [1]. Until one does, the benefit stays in the lab.
• Access: Benefits depend on devices and connectivity. Without them, a digital divide remains.

Way forward

• Record more diverse speakers and build ASR next.
• Tie up with state departments in Sourashtra-speaking areas, following the Sansad Bhashini model [2].
• Publish usage data, not just launch announcements.

Translation is the base of digital inclusion, but voice, representative data, real deployment and connectivity decide whether it reaches people. Government–academia partnerships like DIBD–IISc can be repeated for other languages [1]. Scaled up, they would serve the spirit of Article 29(1), the right to conserve one's language and script, and make Digital India truly multilingual.