The government of India is reportedly working on new regulations for artificial intelligence (AI) companies and generative AI models. The amendments to the Information Technology (IT) Rules are expected to include guidelines to control bias in AI algorithms and language models used for training machines. The Ministry of Electronics and Information Technology is likely to introduce these amendments soon, mandating that AI models with biases or trained on biased datasets be barred from open use. The regulations will also cover deepfake and synthetic content, as well as instructions for platforms dealing with loan apps. The parameters for bias will encompass factors such as caste, religion, community, and national security.
Platforms will be legally obliged to adhere to these parameters, expanding the trust and safety framework of the internet. An anonymous senior government official stated that consultations will be conducted to further eliminate bias and ensure the safety of AI models. Testing processes, including sandbox and stress testing, will be required to detect and rectify biases before the models are released in the market. The urgency to amend the IT Rules may be driven by the rapid progress in generative AI technology and the delayed release of the Digital India Bill, which includes provisions for algorithmic accountability.
The Digital India Bill, initially expected to be introduced sooner, may now be delayed until after the general election. The amendment urgency arises from the rapid progress in generative AI technology. In addition to addressing bias in AI models, the IT Rules amendments aim to provide specific guidelines for deepfake and synthetic content. Furthermore, loan apps will face increased responsibility for preventing fraudulent advertising.
To prevent such loan apps from operating in India, the IT ministry is likely to implement measures such as whitelisting legitimate applications or implementing a banking gateway that blocks transactions through unapproved apps.
It is important to note that the information presented in this article is based on anonymous sources and should be treated as such until official confirmation is received. The government’s proposal to introduce fresh AI rules to control bias in AI algorithms and language models demonstrates its commitment to ensuring a fair and unbiased AI ecosystem in India. The regulation of deepfake, synthetic content, and loan apps further highlights the government’s efforts to safeguard the online space. This move is expected to contribute to the trust and safety framework of the internet, ultimately benefiting the users and stakeholders involved.