A case for a new pronoun for AI

Source– The post is based on the article “A case for a new pronoun for AI” published in “The Hindu” on 22nd July 2023.

Syllabus: GS3- Science and Technology

Relevance: Issues related AI

News- The article explains the issue of use of pronoun for AI tools

What are some facts about chatbots?

Large Language Models are empowering chatbots to engage in conversations like human experts. In some cases, they outperform even the most skilled individuals.

To utilise LLMs effectively, we need to become proficient in using prompts. These prompts can be user-generated questions or instructions provided by software programmers to elicit specific responses from the algorithm.

Why is there a need for a pronoun for AI tools?

There is a shared agreement that misrepresentation of AI’s identity seems manipulative.

Experts propose that restricting AI from using first person pronouns and other human pronouns could decrease cases of AI mistaken identity. It will make it easier to distinguish machine-generated text.

This is crucial because pronouns play a significant role in defining identity in today’s context.

Even ChatGPT acknowledges that granting AI a distinct identity can help clarify its role and prevent confusion with human beings.

AI requires pronouns to establish an identity that is distinct from that of humans. For ethical and security reasons, human beings should know that conversing with a bot.

What are the prevalent practices by big companies having AI based products?

Apple has stereotyped Siri with a feminine-sounding voice, although it offers masculine and gender-neutral alternatives. The aim is to establish an emotional connection with users.

The original default version of Siri unintentionally reflected society’s gender assumptions. It is portraying a preference for a submissive feminine assistant.

Technology companies approach the use of pronouns with caution. Google’s Smart Compose technology, which automatically completes sentences on Gmail, avoids predicting pronouns to prevent unconscious gender biases in the AI model.

What is the way forward?

Regulators should seize the chance to address this issue proactively before conventional AI pronouns become widely established.

To achieve a systematic solution, regulators should collaborate with lexicographers and linguists to establish a standard for major languages from the outset. Pronoun guidelines can then be included in style guides for prompt engineering AI models.