असंगत आउटपुट के बाद एआई मॉडल पर सलाह भेजी गई: आईटी सचिव।

Advisory on AI models sent after inconsistent outputs: IT secretary

The IT secretary sent an advisory on AI models following inconsistent outputs, aiming to address and rectify potential issues and concerns.

Technology
139
08, Mar, 2024

Jivika Chawla

@JivikaChawla

Advisory on AI models sent after inconsistent outputs: IT secretary

On Thursday, Information Technology Secretary S. Krishnan revealed that the IT ministry issued an advisory on artificial intelligence to tech companies due to observed inconsistencies in generative AI tools. The concern arose as these tools altered responses based on the jurisdiction where questions were posed, highlighting a need for rectification. Krishnan referred to the March 1 advisory, emphasizing its applicability solely to significant social media intermediaries—companies with over 5 million users in India.

The advisory specifically targeted eight major social media intermediaries, urging them to disclose the new generative AI tools they introduce to the market. Krishnan explained that certain generative AI tools exhibited variations in their operations across jurisdictions, particularly in responding to sensitive socio-political queries. This discrepancy raised concerns, prompting the government's request for transparency from the identified intermediaries.

Speaking at the ITech Law conference in Bangalore through a video link, Krishnan clarified the limited scope of the advisory, emphasizing its relevance only to a select group of social media companies. The government's proactive stance aims to ensure accountability and consistency in the deployment of AI tools, especially in addressing sociopolitical questions across diverse jurisdictions.