Vineet Sawant, a delivery driver in Mumbai, faced language barriers in his job, navigating the bustling city with limited English skills. "Being on the road is always stressful," he shares, explaining how he struggled to understand delivery instructions primarily in English. This changed when his employer, Zepto, partnered with Reverie Language Technologies, introducing an AI translation service to their app.
Now, drivers like Mr. Sawant can choose from six different languages, including Marathi. "I don't have to guess anymore," he states, noting how this advancement has significantly improved his delivery efficiency.
India boasts 22 official languages and countless dialects, creating a challenge for effective AI communication. Professor Pushpak Bhattacharyya from IIT Mumbai emphasizes that a lack of high-quality data hampers tech development, especially for less common languages. Reverie’s co-founder, Vivekananda Pani, warns that advancements may risk overshadowing unique dialects if not addressed properly.
To tackle these issues, Bhashini, a government initiative that began in 2022, aims to create high-quality datasets and AI models in all 22 languages. Its progress is impressive, with 350 AI models processing over a billion tasks across various public services.
These initiatives promise a future where rural residents will access government services in their native languages, thanks to dedicated AI projects. Furthermore, Kshitij Jadhav’s research in AI programming aims to assist smokers in quitting through tailored conversations, paving the way for personalized advice in multiple languages.
Mr. Sawant's experience reflects this broader trend. Since the implementation of AI translation, he has significantly increased his daily deliveries, retrieving a newfound confidence as his app speaks his language. "It makes us feel like we belong," he concludes, highlighting the profound impact language inclusivity can have on job performance.
