An Indian State attempts to regulate speech and gets it wrong

Screenshot of the article on Tech Policy Press

This piece was originally published on Tech Policy Press on 2025-07-14. An archive of this page can be found here.

Few will contest the observation that India’s information ecosystem is severely polluted. This is likely an outcome of dysfunction across several different layers. Response to this complex, multi-causal dysfunction requires society to pull different levers. One lever that states and political actors, often incentivised by their down desires to control and shape narratives, deploy is legislation that targets ostensibly “misinformation” or “disinformation” (frequently and incorrectly used interchangeably in public discourse), or “hate speech” but in practice tends to have unintended consequences for free speech and expression. The Indian state of Karnataka, where Bengaluru is located, appears to be the latest to join this bandwagon.

Background

In late June Karnataka’s minister for Law and Parliamentary told reporters that the state’s cabinet was going to discuss legislation pertaining to misinformation and “fake news”, and hate crimes and hate speech.

For context, Karnataka is currently administered by the Indian National Congress (INC), the principal opposition party at a Union/Federal Government level. INC had supported a September 2024 Bombay High Court judgment striking down the union government’s amendments to the IT Rules in 2023 which would have required internet services to take action against information (about the union government) flagged by a government-designated fact checking unit. Yet, the INC-led government in Karnataka had pushed ahead with its own fact checking unit proposal in Karnataka earlier in 2024 despite criticism, pushback, and many IFCN-accredited fact checking organisations choosing not to participate.

Interviews by state cabinet minsters in 2023 suggested this unit would be run the Home Department, implying significant law enforcement involvement likely to lead to criminal prosecutions for speech. Back then, Priyank Kharge, state minister for Information Technology and Biotechnology, had sought to downplay criticism of its approach stating that their approach was distinct from the union government’s attempts in that they were not amending laws but using the “current law of the land.” The unit, called Information Disorder Tracking Unit (IDTU), ran a three month trial during a period that coincided with India’s weeks-long general election last year. Of the 18 First Information Reports (FIR) filed, according the Deccan Herald, most were instances targeting INC and its leaders. To my knowledge, other than broad statistics, detailed learnings/findings from the exercise have not yet been discussed publicly. There also appears to have been no detailed explanation from the state government about the shift from a fact checking unit to broader legislative efforts.

Now, a draft version of the “Karnataka Mis-Information and Fake News (Prohibition) Bill, 2025”(“Misinformation Bill”) has been placed in the public domain by the Internet Freedom Foundation (IFF). Digital rights groups have called for its revision/withdrawal and public consultations, as have editorials. In response, Priyank Kharge has attempted to downplay concerns or confusion by referring to “misinformation” about the bill on Twitter/X. In an interview with The Economic Times he claimed that the “11-page document” was not “the draft” but the “opinion of a few people in the law department,” and also clarified that there will be a “wider level of consultations”. It is not clear whether these “wider” consultations will be also be “public” in nature as Mr. Kharge does not appear to have explicitly said so.

A draft version of the “Hate Speech and Hate Crimes (Prevention and Control) Bill, 2025” (“Hate Speech Bill”) is not currently in the public domain (officially released or otherwise) but a number of its provisions have been reported on.

A problematic approach undergirds both bills

The complicated nature of problems plaguing the information ecosystem, first and foremost require a deeper understanding than we currently have. This is true globally but especially so in the Indian context where research is limited. Based on these learnings, responses could focus on the “supply” or “demand” of low quality information. These interventions may be coercive [penalties, arrests], or participatory and voluntary [awareness programmes, research support] in nature. With the misinformation bill GoK has skipped the research and understanding phase and chosen the coercive route through provisions that effectively criminalise speech. This, despite, many scholars explaining why this approach is fraught. This approach is typically weaponised and used against the press and dissenting voices. The Center for International Media Assistance has identified 4 kinds of legal penalties that can impact press freedom, 3 of which (imprisonment, fines, information removal/correction) are in GoK’s version of the Misinformation Bill. It proposes prison terms ranging from 5 to 7 years for speech-related offences, and 2 years for “abetting” them. It specifies that these will be “non-cognisable” (complaints are not required for law enforcement to take action) and “non-bailable”, and goes further to make it harder to seek “anticipatory bail” in these cases. The hate speech bill includes three year sentences for committing and even “unknowingly” abetting offences under the bill. The hate speech bill also includes provisions for imprisonment and fines. SFLC.in contends that the bill’s approach could create a chilling effect on freedom of speech and expression.

The misinformation bill envisions a 6-member body, completely under executive control, whose charter includes a “complete ban on promotion and spread of fake news on the social media platforms,” and is empowered to imprison people for 7 years and even impose fines on them upto INR 1,000,000 (equivalent to approximately 11,500 USD). Not only is this approach impractical, it also creates mechanisms that will be exploited to target people and speech that those in government disapprove of. The bill is also much broader in scope than the union government’s proposed fact checking unit with amendments to the IT Rules in 2023. For all its flaws, the fact checking unit remit and subsequent obligations on intermediaries were purportedly limited to the business of the union government, even if “business” was not precisely defined.

Both bills also attempt to place obligations/penalties on internet services (described later), but it is not clear that they have the authority to do so. Subjects that union and state governments can legislate are designated in three lists, as per India’s constitution - union, state and concurrent. The union list includes subjects that the union government can legislate, the state list contains that states can enact laws for, and the concurrent list has subjects on which both can pass laws. Regulation pertaining to the internet is considered to come under item 31 of the union list. Medianama adds that article 254 of India’s Constitution prevents states from legislating on subjects under the union government’s domain. If these bills are enacted in their present form, it is likely that courts will be called upon to determine their constitutionality.

In February, SouthFirst reported that advisors to GoK were aware that this could pose a challenge, and were considering different approaches. However, GoK has made no attempts to clarify why it went ahead with this option.

Since 2023, GoK has filed over 2000 cases for hate speech, with state’s chief minister recently directing law enforcement to file hate speech cases without complaints (to be clear, there is jurisprudence from India’s Supreme Court to this effect). India’s criminal codes also have provisions that deal with false information, rumours, defamatory speech, incitement, etc. in physical spaces and on the internet. Often, there are challenges are with implementation, enforcement. And, as Gautam Bhatia argued in The Hindustan Times these are also prone to misuse. Karnataka’s government has not clearly articulated how these existing lacunae will be overcome merely by enacting new legislation.

Poor drafting exacerbates their problems

In academic terms “Misinformation” is defined to exclude knowledge of inaccuracy or an intention to cause harm. The misinformation bill, however, ascribes intent in its definition of the term with fuzzy language such as “knowingly or recklessly”. It also attempts to define “fake news”, a term long considered inadequate, distinctly. As Srishti Joshi and Berges Malu point out in ThePrint, it largely fails in this endeavour, but not without adding uncertainty since it explicitly excludes opinions, satire, comedy, sermons, etc., in one case (for misinformation) but not in another (fake news). It also prescribes different penalties treating each as a different category of offence, and limits the six-member body (called “Fake News on Social Media Regulatory Authority”) to “fake news”.

This body is also charged with “prohibiting” information that is “anti-feminism”, “disrespect(s) … Sanatan Symbols and beliefs”, “promoting superstition” and to ensure publication of information based only on “authentic research” on specific subjects like science, religion, history, philosophy and literature with no effort to clarify what any of these could mean especially since neat categorisation in public discourse is rarely simple.

Meghna Bal notes in The Indian Express, that at least a part of the misinformation bill’s Statement on Objects and Reasons appears to have been copied from another bill called “Prohibition of Fake News on Social Media Bill, 2022”. This perhaps explains why the same section refers to sections of the Indian Penal Code which was replaced by the Bhartiya Nyaha Sanhita (BNS) in 2024.

The hate speech bill, meanwhile, reportedly defines intermediaries in a manner close to the definition found in India’s Information Technology Act but without referring to it, creating scope for diverging definitions over time. It also defines harm as “any emotional, psychological, physical, social or economic harm,” without adequate description of what these mean in the context of the bill. An analysis of the full text of this bill may reveal more inconsistencies and shortcomings.

A broader, public consultation may address some of these inconsistencies but their nature, to this author, is reflective of gaps in either intent, capability or both.

Implications for platform governance

Even as the question of whether or not these bills can impose obligations on internet intermediaries remains unanswered, they do contain provisions that will have significant implications for platform governance.

While defining the offence of misinformation, the misinformation bill aims to restrict direct communication or abetting communication of misinformation in the state of Karnataka if it adversely affects “public health, public safety, public tranquility or the conduct of free and fair elections.” This clause is likely extra-territorial in effect since a person in Karnataka can access information posted anywhere on the internet. Communication is also defined broadly in a way that it can cover sharing information with anyone in the state through “… communication device(s), computer resource(s)…”. It is unclear whether it includes end-to-end encrypted communication apps in its ambit, and what the implications may be.

The bill aims to create special courts that have the power to issue directions to intermediaries, publishers, broadcasters, and others to correct or disable access to information as long as legal proceedings are ongoing. Compounding these problematic clauses are the bill’s attempts to place liability on individuals at “companies” (which presumably includes internet intermediaries) for what it deems as direct violations or negligence. The penalties include the possibility of imprisonment, fines, or both. Combined with the complete absence of safeguards and vague powers for the “Fake News on Social Media Regulatory Authority” all of these provisions can lead to onerous or even proactive information restriction obligations on internet intermediaries.

The hate speech bill aims to hold intermediaries liable for “knowingly or unknowingly” disseminating hate speech. An analysis of a leaked version of the bill by Rudraksha Lakra and Nidhi Jha observes that attempts place liability for “unknowingly” disseminating information could go against the Supreme Court of India’s 2015 Shreya Singhal v/s Union of India decision which stated that intermediaries needed to have ‘actual knowledge’ in the form of a court or government order to be held liable. SFLC.in, in its statement, notes that the penalty violating this can include a prison term of up to 3 years, along with a fine.

It is not clear how different intermediaries will interpret and comply with these clauses. But, if both these bills are enacted with similar provisions, their vagueness could create enough of a liability threat that social media platforms may excessively restrict legitimate speech out of caution.

Not just a slippery slope, but an illiberal slide

There appears to be plenty to criticise and question about the specifics of the bills, the institutional design choices, coercive approach and so on. Of greater concern is that the principal opposition party has chosen a highly restrictive and punitive approach at a time when the actions of multiple institutions in the country are jeopardising or question the importance of values such as freedom of speech and expression, and it should be protecting these principles. All this even before making serious attempts to improve their understanding of the complex, intersecting issues and problems at play. The approach also suggests that GoK is either incapable of dealing with the complexity of the issues it is trying to legislate, or does not see the potential for selective/subjective application as problems. Plus, its attempts to minimise near-unanimous criticism by relying on deflection, technicalities and invoking a “but, we’re the good guys” defence are also deeply worrying. Even if these bills fail the test of judicial review (either in parts or completely), this has already shifted the Overton window on speech issues in the country irrevocably towards the illiberal end of the spectrum.