#FactCheck-Mock Drill Video from Kishtwar Falsely Shared as Terror Attack in Jammu & Kashmir

As AI language models become more powerful, they are also becoming more prone to errors. One increasingly prominent issue is AI hallucinations, instances where models generate outputs that are factually incorrect, nonsensical, or entirely fabricated, yet present them with complete confidence. Recently, ChatGPT released two new models—o3 and o4-mini, which differ from earlier versions as they focus more on step-by-step reasoning rather than simple text prediction. With the growing reliance on chatbots and generative models for everything from news summaries to legal advice, this phenomenon poses a serious threat to public trust, information accuracy, and decision-making.

What Are AI Hallucinations?

AI hallucinations occur when a model invents facts, misattributes quotes, or cites nonexistent sources. This is not a bug but a side effect of how Large Language Models (LLMs) work, and it is only the probability that can be reduced, not their occurrence altogether. Trained on vast internet data, these models predict what word is likely to come next in a sequence. They have no true understanding of the world or facts, they simulate reasoning based on statistical patterns in text. What is alarming is that the newer and more advanced models are producing more hallucinations, not fewer. seemingly counterintuitive. This has been prevalent reasoning-based models, which generate answers step-by-step in a chain-of-thought style. While this can improve performance on complex tasks, it also opens more room for errors at each step, especially when no factual retrieval or grounding is involved.

As per reports shared on TechCrunch, it mentioned that when users asked AI models for short answers, hallucinations increased by up to 30%. And a study published in eWeek found that ChatGPT hallucinated in 40% of tests involving domain-specific queries, such as medical and legal questions. This was not, however, limited to this particular Large Language Model, but also similar ones like DeepSeek. Even more concerning are hallucinations in multimodal models like those used for deepfakes. Forbes reports that some of these models produce synthetic media that not only look real but are also capable of contributing to fabricated narratives, raising the stakes for the spread of misinformation during elections, crises, and other instances.

It is also notable that AI models are continually improving with each version, focusing on reducing hallucinations and enhancing accuracy. New features, such as providing source links and citations, are being implemented to increase transparency and reliability in responses.

The Misinformation Dilemma

The rise of AI-generated hallucinations exacerbates the already severe problem of online misinformation. Hallucinated content can quickly spread across social platforms, get scraped into training datasets, and re-emerge in new generations of models, creating a dangerous feedback loop. However, it helps that the developers are already aware of such instances and are actively charting out ways in which we can reduce the probability of this error. Some of them are:

1. Retrieval-Augmented Generation (RAG): Instead of relying purely on a model’s internal knowledge, RAG allows the model to “look up” information from external databases or trusted sources during the generation process. This can significantly reduce hallucination rates by anchoring responses in verifiable data.
2. Use of smaller, more specialised language models: Lightweight models fine-tuned on specific domains, such as medical records or legal texts. They tend to hallucinate less because their scope is limited and better curated.

Furthermore, transparency mechanisms such as source citation, model disclaimers, and user feedback loops can help mitigate the impact of hallucinations. For instance, when a model generates a response, linking back to its source allows users to verify the claims made.

Conclusion

AI hallucinations are an intrinsic part of how generative models function today, and such a side-effect would continue to occur until foundational changes are made in how models are trained and deployed. For the time being, developers, companies, and users must approach AI-generated content with caution. LLMs are, fundamentally, word predictors, brilliant but fallible. Recognising their limitations is the first step in navigating the misinformation dilemma they pose.

References

• https://www.eweek.com/news/ai-hallucinations-increase/
• https://www.resilience.org/stories/2025-05-11/better-ai-has-more-hallucinations/ 
• https://www.ekathimerini.com/nytimes/1269076/ai-is-getting-more-powerful-but-its-hallucinations-are-getting-worse/ 
• https://techcrunch.com/2025/05/08/asking-chatbots-for-short-answers-can-increase-hallucinations-study-finds/ 
• https://en.as.com/latest_news/is-chatgpt-having-robot-dreams-ai-is-hallucinating-and-producing-incorrect-information-and-experts-dont-know-why-n/
• https://www.newscientist.com/article/2479545-ai-hallucinations-are-getting-worse-and-theyre-here-to-stay/
• https://www.forbes.com/sites/conormurray/2025/05/06/why-ai-hallucinations-are-worse-than-ever/ 
• https://towardsdatascience.com/how-i-deal-with-hallucinations-at-an-ai-startup-9fc4121295cc/ 
• https://www.informationweek.com/machine-learning-ai/getting-a-handle-on-ai-hallucinations 

‍

Introduction

‍
The Digital Personal Data Protection (DPDP) Act 2023 of India is a significant transition for privacy legislation in this age of digital data. A key element of this new law is a requirement for organisations to have appropriate, user-friendly consent mechanisms in place for their customers so that collection, use or removal of an individual's personal data occurs in a clear and compliant manner. As a means of putting this requirement into practice, the Ministry of Electronics and Information Technology (MeitY) issued a comprehensive Business Requirements Document (BRD) in June 2025 to guide organizations, as well as Consent Managers, on how to create a Consent Management System (CMS). This document establishes the technical and functional framework by which organizations and individuals (Data Principals) will exercise control over the way their data is gathered, used and removed.

‍

Understanding the BRD and Its Purpose

‍

BRD represents an optional guide created as part of the "Code for Consent" programme run by MeitY in India. The purpose of the BRD is to provide guidance to startups, digital platforms and other enterprises on how to create a technology system that supports management of user consent per the requirements of the DPDP Act. Although the contents of the BRD do not carry any legal weight, it lays out a clear path for organisations to create their own consent mechanisms using best practices that align with the principles of transparency, accountability and purpose limitation in the DPDP Act.

‍

The goal is threefold:

‍

1. Enable complete consent lifecycle management from collection to withdrawal.
2. Empower individuals to manage their consents actively and transparently.
3. Support data fiduciaries and processors with an interoperable system that ensures compliance.

‍

Key Components of the Consent Management System

‍

The BRD proposes the development of a modular Consent Management System (CMS) that provides users with secure APIs and user-friendly interfaces. This system will allow for a variety of features and modules, including:

‍

1. Consent Lifecycle Management – consent should be specific, informed and tied to an explicit purpose. The CMS will manage the collection, validation, renewal, updates and withdrawal of consent. Each transaction of consent will create a tamper-proof “consent artifact,” which will include the timestamp of creation as well as an ID identifying the purpose for which it was given.
2. User Dashboard – A user will be able to view and modify the status of their active, expired or withdrawn consent and revoke access at any time via the multilingual user-friendly interface. This would make the system accessible to people from different regions and cultures.
3. Notification Engine – The CMS will automatically notify users, fiduciaries and processors of any action taken with respect to consent, in order to ensure real-time updates and accountability.
4. Grievance Redress Mechanism – The CMS will include a complaints mechanism that allows users to submit complaints related to the misuse of consent or the denial of their rights. This will enable tracking of the complaint resolution status, and will allow for escalation if necessary.
5. Audit and Logging – As part of the CMS's internal controls for compliance and regulatory purposes, the CMS must maintain an immutable record of every instance of consent for auditing and regulatory review. The records must be encrypted, time-stamped, and linked permanently to a user and purpose ID.
6. Cookie Consent Management – A separate module will enable users to manage cookie consent for websites separately from any other consents.

‍

Roles and Responsibilities

‍

The BRD identifies the various stakeholders involved and their associated responsibilities.

‍

1. Data Principals (Users): The user has full authority to give, withhold, amend, or revoke their consent for the use of their personal data, at any time.
2. Data Fiduciaries (Companies): Companies (the fiduciaries) must collect the data principals' consents for each particular reason and must only begin processing a data subject's personal data after validating that consent through the CMS. Companies must also provide the data principals with any information or notifications needed, as well as how to resolve their complaints.
3. Data Processors: Data Processors must strictly adhere to the consent stated in the CMS, and Data Processors may only process personal data on behalf of the Data Fiduciary.
4. Consent Managers: The Consent Managers are independent entities that are registered with the Data Protection Board. They are responsible for administering the CMS, allowing users to manage their consent across different platforms.

‍

This layered structure ensures transparency and shared responsibility for the consent ecosystem.

‍

Technical Specifications and Security

‍

The following principles of the DPDP Act must be followed to remain compliant with the DPDP Act.

‍

• End-to-End Encryption: All exchanges of data with users must be encrypted using a minimum of TSL 1.3 and also encrypting within that standard.
• API-First Approach: API’s will be utilized to validate, withdraw and update consent in a secured manner using external sources.
• Interoperability/Accessibility: The CMS needs to allow for users to utilize several different languages (e.g. Hindi, Tamil, etc.) and be appropriate for use with various types of mobile devices and different abilities.
• Data Retention Policy: The CMS should also include automatic deletion of consent data (when the consent has expired or has been withdrawn) in order to maintain compliance with data retention limits.

‍

Legal Relevance and Timelines

‍

While the BRD itself is not enforceable, it is directly aligned with the upcoming enforcement of the DPDP Act, 2023. The Act was passed in August 2023 but is expected to come into effect in stages, once officially notified by the central government. Draft implementation rules, including those defining the role of Consent Managers, were released for public consultation in early 2025.

For businesses, the BRD serves as an early compliance tool—offering both a conceptual roadmap and technical framework to prepare before the law is enforced. Legal experts have described it as a critical resource for aligning data governance systems with emerging regulatory expectations.

‍

Implications for Businesses

‍

Organizations that collect and process user data will be required to overhaul their consent workflows:

1. No blanket consents: Every data processing activity must have explicit, separate consent.
2. Granular audit logs: Companies must maintain tamper-proof logs for every consent action.
3. Integration readiness: Enterprises need to integrate their platforms with third-party or in-house CMS platforms via the specified APIs.
4. Grievance redress and user support: Systems must be in place to handle complaints and withdrawal requests in a timely, verifiable manner.

Failing to comply once the DPDP Act is in force may expose companies to penalties, reputational damage, and potential regulatory action.

‍

Conclusion

‍
The BRD on Consent Management of India is a forward-looking initiative laying a technological framework that is an essential component of the DPDP Act concerning user consent; Although not yet a legal document, it provides an extent of going into all the necessary discipline for companies to prepare. As data protection grows in importance, developing consent mechanisms based on security, transparency, and the needs of the user is no longer just a regulatory requirement, but rather a requirement for the development of trust. This is the time for businesses to establish or implement CMS solutions that support this objective to be better equipped for the future of data governance in India.

‍

References

1. https://d38ibwa0xdgwxx.cloudfront.net/whatsnew-docs/8d5409f5-d26c-4697-b10e-5f6fb2d583ef.pdf
2. https://ssrana.in/articles/ministry-releases-business-requirement-document-for-consent-management-under-the-dpdp-act-2023/ 
3. https://dpo-india.com/Blogs/consent-dpdpa/ 
4. https://corporate.cyrilamarchandblogs.com/2025/06/the-ghost-in-the-machine-the-recent-business-requirement-document-on-consent/ 
5. https://www.mondaq.com/india/privacy-protection/1660964/analysis-of-the-business-requirement-document-for-consent-management-system 

‍

Introduction

In an age where the lines between truth and fiction blur with an alarming regularity, we stand at the precipice of a new and dangerous era. Amidst the wealth of information that characterizes the digital age, deep fakes and disinformation rise like ghosts, haunting our shared reality. These manifestations of a technological revolution that promised enlightenment instead threaten the foundations upon which our societies are built: trust, truth, and collective understanding.

These digital doppelgängers, enabled by advanced artificial intelligence, and their deceitful companion—disinformation—are not mere ghosts in the machine. They are active agents of chaos, capable of undermining the core of democratic values, human rights, and even the safety of individuals who dare to question the status quo.

The Perils of False Narratives in the Digital Age

As a society, we often throw around terms such as 'fake news' with a mixture of disdain and a weary acceptance of their omnipresence. However, we must not understate their gravity. Misinformation and disinformation represent the vanguard of the digital duplicitous tide, a phenomenon growing more complex and dire each day. Misinformation, often spread without malicious intent but with no less damage, can be likened to a digital 'slip of the tongue' — an error in dissemination or interpretation. Disinformation, its darker counterpart, is born of deliberate intent to deceive, a calculated move in the chess game of information warfare.

Their arsenal is varied and ever-evolving: from misleading memes and misattributed quotations to wholesale fabrications in the form of bogus news sites and carefully crafted narratives. Among these weapons of deceit, deepfakes stand out for their audacity and the striking challenge they pose to the concept of seeing to believe. Through the unwelcome alchemy of algorithms, these video and audio forgeries place public figures, celebrities, and even everyday individuals into scenarios they never experienced, uttering words they never said.

The Human Cost: Threats to Rights and Liberties

The impact of this disinformation campaign transcends inconvenience or mere confusion; it strikes at the heart of human rights and civil liberties. It particularly festers at the crossroads of major democratic exercises, such as elections, where the right to a truthful, unmanipulated narrative is not just a political nicety but a fundamental human right, enshrined in Article 25 of the International Convention on Civil and Political Rights (ICCPR).

In moments of political change, whether during elections or pivotal referenda, the deliberate seeding of false narratives is a direct assault on the electorate's ability to make informed decisions. This subversion of truth infects the electoral process, rendering hollow the promise of democratic choice.

This era of computational propaganda has especially chilling implications for those at the frontline of accountability—journalists and human rights defenders. They find themselves targets of character assassinations and smear campaigns that not only put their safety at risk but also threaten to silence the crucial voices of dissent.

It should not be overlooked that the term 'fake news' has, paradoxically, been weaponized by governments and political entities against their detractors. In a perverse twist, this label becomes a tool to shut down legitimate debate and shield human rights violations from scrutiny, allowing for censorship and the suppression of opposition under the guise of combatting disinformation.

Deepening the societal schisms, a significant portion of this digital deceit traffic in hate speech. Its contents are laden with xenophobia, racism, and calls to violence, all given a megaphone through the anonymity and reach the internet so readily provides, feeding a cycle of intolerance and violence vastly disproportionate to that seen in traditional media.

Legislative and Technological Countermeasures: The Ongoing Struggle

The fight against this pervasive threat, as illustrated by recent actions and statements by the Indian government, is multifaceted. Notably, Union Minister Rajeev Chandrasekhar's commitment to safeguarding the Indian populace from the dangers of AI-generated misinformation signals an important step in the legislative and policy framework necessary to combat deepfakes.

Likewise, Prime Minister Narendra Modi's personal experience with a deepfake video accentuates the urgency with which policymakers, technologists, and citizens alike must view this evolving threat. The disconcerting experience of actor Rashmika Mandanna serves as a sobering reminder of the individual harm these false narratives can inflict and reinforces the necessity of a robust response.

In their pursuit to negate these virtual apparitions, policymakers have explored various avenues ranging from legislative action to penalizing offenders and advancing digital watermarks. However, it is not merely in the realm of technology that solutions must be sought. Rather, the confrontation with deepfakes and disinformation is also a battle for the collective soul of societies across the globe.

As technological advancements continue to reshape the battleground, figures like Kris Gopalakrishnan and Manish Gangwar posit that only a mix of rigorous regulatory frameworks and savvy technological innovation can hold the front line against this rising tidal wave of digital distrust.

This narrative is not a dystopian vision of a distant future - it is the stark reality of our present. And as we navigate this new terrain, our best defenses are not just technological safeguards, but also the nurturing of an informed and critical citizenry. It is essential to foster media literacy, to temper the human inclination to accept narratives at face value and to embolden the values that encourage transparency and the robust exchange of ideas.

As we peer into the shadowy recesses of our increasingly digital existence, may we hold fast to our dedication to the truth, and in doing so, preserve the essence of our democratic societies. For at stake is not just a technological arms race, but the very quality of our democratic discourse and the universal human rights that give it credibility and strength.

Conclusion  

In this age of digital deceit, it is crucial to remember that the battle against deep fakes and disinformation is not just a technological one. It is also a battle for our collective consciousness, a battle to preserve the sanctity of truth in an era of falsehoods. As we navigate the labyrinthine corridors of the digital world, let us arm ourselves with the weapons of awareness, critical thinking, and a steadfast commitment to truth. In the end, it is not just about winning the battle against deep fakes and disinformation, but about preserving the very essence of our democratic societies and the human rights that underpin them.