Tech industry to push back against the new AI rule requiring govt approvals
The Minister of Electronics and IT’s diktat that makes it mandatory for tech companies working on artificial intelligence products to seek government nod before launching their products in India has attracted backlash from the industry and legal experts.
- Also read: Effective use of technology will create hunger for talent: WEF MD
“It’s an advisory that is not based on law but is a trailer of what kind of regulations we can expect from the Ministry of Electronics and Information Technology (MeitY) in the future. It’s not legally binding, so I can not tell what section or what law MeitY is asking companies to submit a report within 15 days. If the labeling of outputs of generative AI is the only requirement, one can see the compelling argument in its favour. If, however, the idea is to seek prior permission before any launch, that’s unnecessary interference which innovators can live without,” said Mishi Choudhary, technology lawyer and online civil liberties activist.
MeitY issued a notice on March 1 warning platforms that failing to comply could result in legal action.
Essentially, AI platforms need to label under-trial AI, seek approval from the government before deploying AI models deemed “under-testing” or “unreliable”, and obtain explicit consent from users before exposing them to such AI models.
Experts on the other hand argue that this advisory has no legal basis in the IT rules. An executive working in the policy department of a large tech companies, who did not wish to be named, told businessline, “Recently the government has started to extend the powers of the due diligence rules of the IT rules indiscriminately. This is just another example of that. This advisory is untenable and impractical on multiple levels, if it challenged in court it will be thrown out. Moreover since it is an advisory, we are now left wondering if it is mandatory, the government has been extremely unclear.”
‘Lacks clarity’
The view that this advisory lacks clarity was echoed by Vivan Sharan, Partner at Koan Advisory Group, an India focused public policy firm, which works with tech firms. Sharan explained to businessline, “The Centre has not laid out the testing threshold for these AI firms. Any company can say that their tools are thoroughly tested for deployment amongst the consumer. This is a ad hoc response these standards need to be discussed and defined before issuing such an advisory in the first place.”
- Also read: Perfect promises, imperfect realities of AI
The tech executive quote above added that for a technology like AI, which is evolving and changing rapidly, a nuanced technical discussion needs to be conducted before regulations are decided. “The Centre is overreaching with its regulations, as a industry we are still understanding what bias in AI, dissemination of wrongful by AI actually means and how it can corrected. An AI bot is an aggregator of information which is already in circulation. Technically, individuals disseminating this information need to be held accountable if the Centre wants accountability. And we have legal systems in place for that.”
However, Abhishek Malhotra, a prominent lawyer in the tech sector believes that there is some legal basis to issue this advisory. It is nebulous but, “The current announcement comes in the form of an advisory which is not, per se, binding. However, if not followed, the offer of such AI models may be proscribed either through an amendment or by deploying the provision that allows blocking orders to be issued under Section 69 of the IT Act, read with Rule 3 of the IT Rules that require the intermediaries to not engage in an action that is in violation of Articles 14, 19 and 21 of the constitution of India,” he said.
SHARE
Published on March 3, 2024
