Meta is embarking on a new strategy to enhance its AI-powered chatbots by hiring US-based contractors to develop character-driven interactions tailored for key international markets, including India, Indonesia, and Mexico. According to job descriptions reviewed by Business Insider, the company plans to pay these contractors up to $55 per hour for their fluency in Hindi, Indonesian, Spanish, and Portuguese.
The roles are focused on providing creative direction and shaping the personalities of chatbots across Meta’s platforms, such as Instagram, Messenger, and WhatsApp. The aim is to tailor these AI companions to the specific languages and cultural nuances of each target market outside the United States. Candidates are expected to possess at least six years of experience in storytelling and character creation, coupled with a strong understanding of prompt engineering and AI content pipelines.
Meta has not yet released an official statement regarding these hiring initiatives. However, CEO Mark Zuckerberg has previously highlighted the potential of these chatbots to broaden the company’s social influence. He envisions them as tools that could complement real-world friendships and foster a greater acceptance of digital companionship. “Over time, we’ll find the vocabulary as a society to be able to articulate why that is valuable,” Zuckerberg stated in a recent podcast.
Meta’s initial foray into AI-powered chatbots began in 2023 with celebrity alter-ego characters, featuring personalities like Kendall Jenner, Snoop Dogg, and Tom Brady. The company invested millions in these celebrity-backed virtual personas. However, within a year, Meta shifted its strategy, discontinuing the celebrity bots and introducing a system that allows users to create their own chatbots.
In 2024, Meta launched AI Studio, a toolkit that empowers anyone to design and build chatbots. AI Studio is currently available in the US and Indonesia, and platforms like Instagram, WhatsApp, and Messenger host numerous character-driven chatbots created by both influencers and everyday users. This shift towards user-generated content allows for a more diverse and personalized chatbot experience.
By hiring contractors to create localized chatbots, Meta is signaling a more direct involvement in shaping the personalities of these AI entities, ensuring they resonate with audiences in countries like India and Indonesia. This hands-on approach contrasts with relying solely on creator-built bots for growth in these markets.
Meta’s increasing investment in chatbots has also raised concerns about potential misuse and ethical implications. A Reuters investigation revealed that Meta’s internal guidelines had previously allowed its bots to engage in inappropriate conversations with teenagers and children, disseminate inaccurate medical information, and generate racist content. These findings led to calls from US lawmakers for greater scrutiny of Meta’s AI policies, with senators demanding investigations into the development and monitoring of these bots. While Meta has since modified its rules, the incident underscores the inherent risks of deploying character-driven AI chatbots on a large scale.
Privacy concerns further complicate the issue. A Business Insider report in August revealed that contractors reviewing Meta’s AI conversations regularly encountered personally identifiable information (PII), including users’ names, emails, phone numbers, and even selfies. This exposure highlights the potential for sensitive user data to be compromised when human reviewers are involved. Meta has stated that it enforces strict policies and technical safeguards to protect user data.
In the US, some of Meta’s more controversial chatbot personas included “Russian Girl,” “Step Sister,” and “Step Mom,” which users discovered could engage in overtly sexual roleplay. The notoriety of these bots became so widespread that an OpenAI employee publicly criticized Meta AI chief Alexandr Wang for overseeing a system that featured such characters.
In Indonesia, Business Insider found that two of the most popular Meta AI characters were named “Lonely woman” and “Deviant male.” While conversations with these characters remained light-hearted and non-sexual in Business Insider’s tests, their names and popularity highlight the delicate balance Meta and its contractors must maintain between user appeal and reputational risk.
Meta is utilizing staffing vendors to fill these new roles. Crystal Equation, for example, had job listings for Hindi and Indonesian language positions explicitly stating the roles were for Meta. Aquent Talent advertised Spanish-language roles for a “top social media company” based in Menlo Park, focused on building “technologies that help people connect.” Neither firm responded to requests for comment from Business Insider.
This initiative reflects Meta’s commitment to expanding its AI capabilities and tailoring its services to diverse global audiences, while also navigating the ethical and privacy challenges that come with deploying advanced AI technologies.
