You can earn ₹5,000 an hour by building a Hindi AI chatbot for Meta: Get all the details

Meta is ramping up efforts to develop AI-powered chatbots with a local flavour, hiring contractors at rates of up to $55 (around Rs 5,000) an hour to design digital personalities for its platforms. According to job postings reviewed by Business Insider, the roles are focused on creating characters in Hindi, Indonesian, Spanish, and Portuguese for use across Instagram, Messenger, and WhatsApp.

The company is seeking applicants with at least six years of experience in storytelling, character creation, and a working knowledge of prompt engineering. Unlike traditional coding roles, these jobs prioritise creativity and cultural understanding, ensuring that AI companions resonate authentically with users in markets like India, Indonesia, and Mexico. Staffing agencies, including Crystal Equation and Aquent Talent, have been advertising the positions on Meta’s behalf.

Meta has experimented with AI-driven characters before. In 2023, it launched celebrity alter-egos played by Kendall Jenner, Snoop Dogg, and Tom Brady, an expensive experiment that was quietly wound down within a year. This was followed in 2024 by AI Studio, a toolkit that allowed creators and influencers to design their own chatbots, hundreds of which now populate Meta’s apps. But the new wave of contractor hires suggests the company wants more control over how these digital personas are shaped.

CEO Mark Zuckerberg has argued that chatbots could one day complement real-world friendships, becoming a natural part of daily interaction. Yet, the push into AI personalities has not been without controversy. A Reuters investigation found that Meta’s bots had previously flirted with teenagers, dispensed questionable medical advice, and even produced racist content. The revelations prompted US lawmakers to call for tighter oversight of the company’s AI practices.

Privacy issues have also come under scrutiny. A Business Insider report last year revealed that contractors reviewing AI conversations often stumbled upon personal user data such as names, phone numbers, emails, and even selfies. While Meta has insisted it enforces strict safeguards, the risks of leaks remain a concern.

Adding to the challenges are the sometimes controversial personas themselves. In the US, chatbots with names like “Russian Girl” and “Step Mom” gained notoriety for enabling sexually suggestive roleplay, while in Indonesia, characters such as “Lonely woman” and “Deviant male” attracted significant engagement. These examples underline the fine balance Meta must strike between offering engaging experiences and avoiding reputational pitfalls.

For now, the company is betting on professional writers and cultural experts to craft characters with just the right mix of relatability and local nuance. Whether this strategy helps Meta normalise AI companions or triggers new rounds of criticism remains to be seen.