- This topic has 3 replies, 4 voices, and was last updated 1 year, 4 months ago by
.
-
Posts
-
-
UThe main intention was to make a politically unbiased language model, but it didn’t go as planned. Imagine it as deepseek/chatgpt which can think out, without any predefined ideological alignment.
We were very close to launching it, but backed off at the last moment while realising the potential consequences. If we restart to fix those it’ll cost us a lump of money (which atleast for us, is a big amount).
By violet acts, I mean instructions to built bombs, procedures to synthesize lethal drugs etc.
What should we do?
-
-
-
UThe Digital India Act will cover most of these regulatory issues with AI once it replaces the IT act. As of now, there is no regulation or law that governs AI in India.
However, that doesn’t mean you can do whatever you wish with your AI model. There is still the requirement of registering it with MeitY meanwhile, which explicitly prohibits the use of these models for unlawful/illegal purposes.
https://www.lexology.com/library/detail.aspx?g=1995b9a6-d3c9-44f2-b9b8-4f05d69bfc18
There are enough provisions under current cybercrime laws to tackle these issues as and when they happen. It’s not worth the risk. Even if you use a disclaimer or reproduce existing information available on the internet, you cannot claim safe harbour because you’re using an Al model that will recreate its own interpretation of the information. So you are liable for any adverse consequences.
-
