Sign In
By signing in or creating an account, you agree with Associated Broadcasting Company's Terms & Conditions and Privacy Policy.
By signing in or creating an account, you agree with Associated Broadcasting Company's Terms & Conditions and Privacy Policy.
New Delhi: The draft copyright framework of artificial intelligence in India can permit the use of almost any publicly available online material to train AI models, unless the platform provides active blocking. The discussion is a new working paper on the Department of Promotion of Industry and Internal Trade (DPIIT) suggests a system where legally obtained material should be considered fair by developers, and it is up to the sites to put paywalls or technical barriers in place in order to protect their content.
Experts say that this model is a significant change in the way copyright operates in cyberspace. It is also a good way of increasing the content one can use to train unless it is restricted by access control. Lawyers have cautioned that this may put a strong weight on the platforms and creators who will have to spot the violations and handle anti-scraping tools to maintain the command of their work.
The offer is called 'One Nation, One Licence, One Payment: Balancing AI Innovation and Copyright', which proposes a mandatory licensing system under a new Copyright Royalty Collection and Allocation Tribunal (CRCAT). With this system, the determining factor of training rights is accessibility and not explicit permission. The content that is maintained on freely accessible platforms would automatically be used to train AI.
Legal scholars observe that this is contrary to the conventional perception of the copyright, which is unprotected public-viewable material unless licensed. They caution that the operational and technical demands of the installation of robust access controls impose burdens on platforms that choose not to conform.
According to policy analysts, the proposed regime can also pose a problem to AI companies. The draft model would make global revenue and royalty rates set by governments and the tribunal structure be reported annually, and an annual report might not reflect smaller creators entirely. This may result in protracted audits and disagreements on royalty payment and governance issues.
Critics claim that no global framework has ever demonstrated that mandatory AI training licences provide any significant income to individual innovators. Historical statutory licensing patterns tended to yield poor or disproportionate payouts, which casts doubt on whether the new model would serve the interests of smaller artists or writers in India.
Along with sites that decide to censor its content, it will rely on its capability to identify scraping or unlicensed applications. Lawyers explain that it is a challenge to keep an eye on violations of the law on a large scale, and the absence of powerful detection mechanisms will not allow platforms to retain control of their activities within the AI training ecosystem.
