Female bots, male platforms reveal gender bias in AI
Companies turn to female personas for customer service but give male names to tech platforms
What do Eva, Ruhh, Anya, Keya, Disha, and Sia have in common? If you thought they were girls with fancy names, you are not too far from the truth. They are chatbots — a program or artificial intelligence that conducts conversations — that interact with customers of companies across different sectors. ROMITA MAJUMDAR writes
What do Eva, Ruhh, Anya, Disha, and Sia have in common? If you thought they were girls with fancy names, you are not too far from the truth. They are chatbots— a program or artificial intelligence( AI) that conducts conversations—that interact with customers of companies across sectors such as banking, insurance, pharmaceutical, and tourism.
Indian companies are not lagging their global counterparts, like an Alexa or Siri, in introducing chat bots to ease customer service.
However, while the bots are female, companies choose masculine names for their technology platforms. Except Infosys’s NIA, technology platforms have names such as IBM Watson, Salesforce Einstein, Wipro Holmes, and TCS Ignio. These are AI or analytics based platforms to assist decisionmaking. “Voice mails, G PS( Global Positioning System) directions, railway announcements to‘ busy’ messages on phones—we hear automated female voices ,” said Sachin Jaiswal, chief executive officer (CEO) and co-founder, Niki.ai. The Bengal ur u-based company has put out many of the chat bots being used now in the Indian market.
He added, “This hints at how we have unconsciously incorporated gender stereotypes, and influence the market. Brands go with habitual flow of human perceptions.” Jaiswal also said people were more likely to talk to bots with an identifiable gender.
Almost all companies are now investing in automation, AI, and digital technology in B 2 Band B 2 C operations. On the consumer facing side, the investment is usually in a chatbot.
These take queries, offer suggestions and new products, and collect feedback. These tasks were performed earlier by call centre executives, receptionists, and help desk staff. Most of these jobs went to women.
There are some technology reasons as well, saide xperts.
The range of the adult female voice is 90 Hz, higher than the male voice (70 Hz ). This makes it easier to syn the si se. In other words, if a company wants to convert their text-based chat bot into a voice-based one, a female voice will be more audible.
For enterprise technology, which are not always voice-based, there is no such requirement.
Sanjeev Nair, co-founder and chief operations officer of Light Information Systems said the female voice often came across as more empathetic.
“If the bot cannot help the user with the issue at hand, the overall experience of the system will be rated low, irrespective of the avatar of the bot. Personas are a novelty when things are new— it helps with on-boarding of users by giving them a sense of relatability of interacting with a human-like system, but (gender) has limited impact on long-term adoption,” Nair added. Light Information Systems specialises in creating conversational chatbots.
Experts said the market was still nascent, and overtime, consumers were likely to lose the need for gendered chatbots. That, however, does not explain why companies continue to choose masculine names for their AI or analytics platforms.