LSPR Monthly Newsletter (Volume 8)

The Next Big Reform: Why the Government of India Should Move from Cash Accounting to Accrual Accounting

Parv Tyagi

“In Greece, statistics is a combat sport.” Andreas Georgiou declared in 2011, after the revered statistician was slapped with criminal charges. His crime – reporting correct budget deficit. Georgiou was appointed, in 2010, as the head of ELSTAT, Greece’s statistics agency, after inquiries by the IMF and the European Commission suggested underreporting of budget deficit. In early 2009, Greece’s debt laden government forecasted a lowly 3.7 percent budget deficit. Georgiou revised it upwards to 15.8 percent! The brazenness of the fiscal fraud was astonishing. As was the witch hunt that followed.

In any case, creative accounting is not new to governments – least of all to the Indian government. The true extent of debt is often pushed under the carpet by deferring cash disbursements and extra-budget borrowings such that they do not reflect in the budget. In 2021, however, the Union budget took a break from the past and discontinued the practice of extra-budget borrowings. This disontinuation has sadly not continued into the recently announced budget for FY 2023-24.

Transparency in Off-Budget Financing

The most sought-after number in a government budget is fiscal deficit. In simple words, it is the difference between a government’s total revenue and its total expenditure, expressed as a percentage of GDP. The Union Minister of Finance Nirmala Sitharaman in the budget for FY 2022-22 pegged the fiscal deficit for the previous FY 2020-21 at 9.5%, up from 3.5% budgeted the year before. A near collapse of tax and non-tax revenue, besides increased spending during the pandemic caused the number to shoot up. However, the deficit widened also because the minister decided to discontinue the National Small Saving Fund (NSSF) loan to Food Corporation of India (FCI), thereby bringing the off-budget food subsidy bill on-budget. Let’s understand the math.

A nodal agency of the union government, the Food Corporation of India buys wheat and paddy from farmers at a set minimum support price. It then distributes it, through the public distribution system, to NFSA beneficiaries at highly subsidised prices. The central government has to reimburse the FCI this difference in procurement price and subsidised price in form of food subsidy. The allocations made to food subsidy in the budget, however, used to be far less than the actual costs incurred by the FCI. The FCI would meet this shortfall, or a bulk of it, by borrowing from the NSSF every year.

Put simply – instead of borrowing itself, the govt would direct the concerned public institution to take loans and pay for its expenditure. Since no cash left the government’s coffers in the process, the transaction would not get recorded in the budget.

The Indian government long managed to keep its fiscal deficit number artificially low by such off-budget borrowings. This is made possible because unlike private entities, Government of India follows cash accounting. It records a transaction only when cash changes hands. So, expenditure is recorded when cash is paid out and revenue is recorded when cash is received in.

For instance, the headline fiscal deficit for the year 2018-19 was reported to be 3.4%. The actual number after accounting for off budget borrowings, however, was 4.26%. Similarly, against the headline fiscal deficit of 4.6% in 2019-20, the actual fiscal deficit was 5.32%.

It is this practice of deferring expenditure off the books, that the Finance Minister discontinued two years back.

It may be important to note that this accounting jugglery costed the exchequer. Mere postponement of payment does not extinguish the govt’s liability to pay these dues. Instead, by deferring payments, the govt. would incur additional cost by way of interest obligations. Also, financing from high interest credit facilities like NSSF would raise the cost of subsidies on the whole. For reference, government bond yields (the rate at which govt can directly raise money in the bond market) usually float at around 6 percent, as against NSSF interest rates of 8-9 percent (or more). Worse still, it makes govt accounts less transparent. The CAG report on Compliance of the FRBM Act, 2003, tabled in 2019 indicted the government of concealing deficits in accounting finesse.

By discontinuing off budget NSSF loans to FCI, the Finance Minister made govt accounts more transparent. The divergence between headline and actual fiscal deficit reduced in 2020-21 and reduced further in 2021-22. However, the Minister reversed this good practice in this year’s budget, and food subsidy may again have to be funded off-budget, like the past.

The subsidy on food for 2023-24 is estimated to be ₹1.97 lakh crore, compared to ₹2.87 lakh crore in 2022-23. The expenditure profile of the Budget 2023 shows that the FCI is provisioned to borrow ₹15,000 crore in the current fiscal year; ₹40,000 crore through external commercial borrowings (ECB) and another ₹1.05 lakh crore through “others” (read NSSF) in the next fiscal year. In short, off budget borrowings are back!

This points to the simple conclusion that the system actually requires quite an overhaul. Fundamental structural changes need to be made to the way public accounts are prepared in India. In this regard, a transition from cash-based accounting to accrual-based accounting ought to have been the logical next step to discontinuation of off budget loans. A transition to accrual numbers would go a long way in solving longstanding problems of non-transparency in government accounts.

Moving Towards Accrual Accounting

The preparation of government accounts in India is governed by the Government Accounting Rules, 1990 framed under Article 150 of the Constitution. As per the rules, the government is to keep its accounts on a cash basis.

The principal objective of accounting is to keep a simple and reliable record of financial transactions. Cash accounting in public sector fails on both these counts. As noted, governments following this system keep a record only of their cash transactions. This limitation is exploited by either rolling over expenditure to the next year or transferring liabilities to third parties (like the FCI in the above case). By contrast, accrual accounting records a transaction as soon as it takes place, regardless of actual receipt or payment of cash. Non-cash assets like land and liabilities like payment arrears are also recorded. Also, transfer of liabilities to third parties (such as public corporations, nodal agencies) is contained as financial statements prepared under accrual system consolidate all entities under govt control. Moreover, regular revaluation of such assets and liabilities makes sure that these stocks are recorded at their market value as opposed to book value. This ensures a truer reflection of government’s financial position at a given point of time. In short, accrual numbers are useful to understand longer term implications of costs incurred today but not payable in short term.

This year marks twenty years since the Twelfth Finance Commission first recommended a changeover to accrual accounting. The recommendation has been reiterated by every Finance Commission since then, including the Fifteenth Finance Commission which submitted its report in 2021. Fifty-seven governments across the globe now account on accrual basis. This gives all the more reason to move away from archaic cash accounts. Adoption of accrual accounting will enhance readability of govt accounts to investors and researchers and increase international acceptability of India’s numbers.

This transition, though, is by no means an easy task. Most countries took more than 10 years to complete the process that requires big investments in IT infrastructure and training of accountants. The absence of professional accounting personnel in India’s public sector is a problem in its own right. But taxpayers’ money cannot be held hostage to small logistical hitches. A sound govt accounting system is about institutional integrity and efficiency. And it is high time the government bites the bullet and undertakes this reform.

The author is an undergraduate student at the National Law School of India University and the current Managing Editor at LSPR.

Unpacking the Pitfalls: A Critical Analysis of Draft IT Rules for Online Gaming

Abhishek Gupta


Recently, the Ministry of Electronics and Information Technology [“MeitY”] introduced further amendments to the IT Rules, 2021 regarding online gaming after it was notified as the nodal ministry for online gaming. The proposed rules require gaming companies to be part of a self-regulatory body, only publish games approved by such bodies, follow know-your-customer (KYC) norms, set up a grievance redressal system, and classify online gaming platforms as intermediaries among other things. However, the proposed rules, which have been opened for public consultation have raised some important concerns like ambiguity in definition of online games, classification of gaming platforms as intermediaries, lack of distinction between games of skill and games of chance, concerns over self-regulatory body and excessive powers in the hands of government.  This article offers a synopsis of these issues which need to addressed by the government before finalising the rules.

Need for Central Regulation of Online Gaming

The current archaic laws have created inconsistency in online gaming regulation. Gambling comes under the State List and has been banned by many states. In the view of many gambling related suicides due to addiction and financial losses, many states have tried to regulate or outright ban the ‘game of chance.’ However, the problem arose when the states banned games that held legal by the Courts called ‘game of skill.’ So, it has been a consistent demand from the gaming industry to clearly demarcate game of skill and game of chance to streamline the process. Additionally, Intra-state activities common in online gaming would not come under the state list since only parliament can make laws having extra-territorial applicability. These concerns stress for a central regulatory framework. However, concerns remain whether the central government lacks legislative competence to regulate online gaming.

Issues with the Draft IT Rules

There are several concerns over the draft IT rules on online gaming:

  1. Definition of Online Games

The proposed definition of online games in the regulations is vague and creates more confusion for the industry. Rule 2(1)(qa) defines “online game” as a “game that is offered on the Internet and is accessible by a user through a computer resource if he makes a deposit with the expectation of earning winnings”. It is unclear whether the free-to-play games that do not require any money in order to participate in the game but have in-game purchases will fall within the purview of the term ‘online gaming.’ Another concern is that the definition of online game seems to legitimise gambling as a player who is gambling makes a deposit with the expectation of winning. The question, then arises whether this makes gambling an online game and thereby legal.

2. The classification of Online Gaming Platforms classified as an Intermediaries

The draft rules classify online gaming platforms as ‘intermediaries.’ According to Rule 2(1)(qb), “online gaming intermediary” means an intermediary that offers one or more than one online game. However, classifying gaming firms as intermediaries makes little sense as most of them are publishers as they are publishing their own content through these games and hence, they are responsible for the content they publish. Essentially, gaming platforms are publishers. Previously, the Ministry of Information & Broadcasting has stated that gaming platforms should be classified as ‘publishers’ and not as an ‘intermediary.’ Section 79 of the IT Act states that an intermediary shall not be liable for hosting any third-party content. Therefore, if online gaming platforms are classified as intermediaries, they will be exempt from penal action for publishing any third-party content. However, they will be held responsible for such content as a publishers.

3. No distinction between Game of Skill and Game of Chance

It was anticipated that the rules will define and make clear demarcation between a game of skill and a game of chance. However, the rules do not define the two terms. The existing gambling laws [Public Gambling Act, 1867] are vague and outdated. They merely state that games of skill are not prohibited under the act. The Act does not define what constitutes these games, and leaves the same for the courts to interpret. This lack of clear distinction has often led to skill-based games getting confused for chance-based games and getting banned. The rules should provide a safe harbour to games of skill. It misses an opportunity to clearly define and distinguish between these two. Additionally, the games of skill and games of chance are taxed differently. While skill-based games are taxed at 18%, chance-based games are taxed at 28%. A clear distinction between the two will make it easier to identify the rate of tax to be levied over such games and help in a good tax regime.

4. Concerns over Self-Regulatory body

The draft IT rules propose the establishment of a Self-regulatory body (SRB) that would be responsible for registering and approving games as well as providing a grievance redressal mechanism. The SRB has also been tasked to come up with a regulatory framework which will include parameters to adjudge and regulate the content of online games and include safeguards against potential harms. An online gaming intermediary will be able to host a game only if a game has been registered and approved by the SRB. The issue that arises is that the wording used in the rules provides discretionary powers to the SRB whether to register a game or not. The phrase ‘may register’ should be replaced with ‘shall register’ to remove the discretion allotted to SRB in registering a game if it fulfils other

conditions in the rules. Furthermore, the decision of the SRB in case of a grievance redressal is final and there is no appellate body to safeguard game publishers if SRB chooses not to register a game. Considering that these SRBs are to be appointed by MeitY, it will likely have influence over such SRB. It will essentially give government broad powers regarding registration of games. Therefore, it is necessary that an appellate body be set-up to prevent such unfair practices by the government or the SRB.

5. Excessive powers to government by Rule 6A

The proposed rule 6A allows the government to declare any game as an online game even if the game does not require any deposit if the government is convinced “that such game may create a risk of harm to the sovereignty and integrity of India or security of the State or friendly relations with foreign States or public order, on account of causing addiction or other harm among children.” This essentially grants powers unbridled to the government creating scope for its possible misuse. The government may block or restrict access to any game it does not like.

Way Forward

The rules are a step in right direction and have largely received a positive response from the gaming industry. However, some issues still persist which need to be addressed

The government needs to ensure that there are no unnecessary hassles for a gamer in order to promote the industry. Rule 4A introduces a know-your-customer (KYC) procedure to be followed by the intermediary for registration of the account of a user. However, most of the times in real money games, the users prefer to play ‘free games,’ where KYC procedure might not be required and will only create hassle for a user. It is also imperative that government supports small, emerging players in the industry and give them due consideration while framing the rules. The Requirement of having three separate employees-Grievance Officer, Chief Compliance Officer and Nodal Contact Officer on the payroll of Online Gaming Intermediary is onerous especially for startups and midsized gaming companies. It is suggested that intermediaries be given the option to have only one employee that may perform all the three functions. It is also proposed that MeitY puts a cap on the fee charged by SRB for registration as it has been noticed that SRBs have charged exponentially high rates for registration. It gives an unfair competitive advantage to large companies as such high fees cannot be afforded by startups and small companies. These changes are necessary for promoting the gaming industry. It was anticipated that the rules will lay a framework for clear and uniform laws across the country. However, the rules are far from achieving it. The government will fare better by consulting various stakeholders to remove grey-areas and address other issues in the proposed rules before finalising it.

Abhishek Gupta, undergraduate student at the National Law University Delhi.

Analysing the French Retirement Reforms in the Indian Context

Oorja Newatia, Kanishk Srinivas

France is in turmoil once again facing waves of nationwide strikes. This time the cause is a reform to its retirement and pension system that the working class sees as an attack on the existing social security structure. In this article, the authors describe the French recruitment reform and analyse the challenges and questions it poses to the Indian pension system.

The article is divided into three parts. First, we evaluate the new French retirement policy, its objectives and the reasons for massive protests against it. Second, the issue of life expectancy linked to retirement age is examined in the Indian context of the increasing workforce and pension bill. Finally, we attempt to provide some solutions that would preserve the social security of pensions without impinging on fiscal stability and employment opportunities.

The French Fiasco

On January 10, the French Government announced a radical change to the country’s existing pension system. The new system envisages increasing the retirement age to 64 years from the existing 62 years through a staggered raising of the retirement age by 3 months every year till 2030. The new scheme also increases the minimum number of years a person needs to work to be eligible for state-funded pension to 43 years starting from 2027.

The policy has been motivated primarily by three factors. First, the French pension system has allegedly become “unsustainable” with pensioners fast outnumbering employees. This is a consequence of the ageing population and declining workforce which – in a system relying on employees’ contributions to fund pensions – is disastrous. There is also a substantial burden on the exchequer with President Macron stating that the current pension system would collapse in the absence of government subsidies. Second, the policy aims to link the increasing life expectancy with retirement age. With healthier lifestyles and better healthcare facilities, the life expectancy in France has increased to 82 years in 2020. It is believed that an increase in life expectancy necessitates that people remain in the workforce longer than they usually did and continue contributing to productive economic activities. Third, the policy is partly influenced by global trends with multiple countries around the world, including France’s neighbouring countries raising their retirement ages and linking it with life expectancy to offset the consequence of smaller workforces.

While the policy seems to be driven by pragmatic concerns of reducing government deficit, French citizens, especially the working class, are opposing it with demonstrations, protests and blockades being

frequently organised. In a society that significantly values the freedom to work and sell services for historical reasons, a scheme compelling them to work longer than usual and compromising the social security framework is unacceptable. The erosion of the social security system has led to dissatisfaction among workers who argue for increasing pension costs to be met out of other sources like increased employers’ and employees’ contributions. Women workers have been at the forefront of protests, arguing that any increase in retirement age may disproportionately impact their ability to secure pensions since they are more likely to take career breaks for family commitments.

With President Macron unwilling to scrap the policy and instead deciding to provide concessions to some categories of workers, a broader debate has opened up about factors influencing retirement age and the structuring of state-funded pension programmes. 

The Indian Perspective

The demand for increasing the retirement age was raised by the Employee Provident Fund Organisation (EPFO) because India is predicted to become an ageing society by 2047 with around 140 million people expected to be above 60 years. The EPFO called for the retirement age to be linked with life expectancy as is being done in other countries. It was argued that raising the retirement age would imply a deposit of higher quantum pensions for a longer duration with EPFO and help offset inflation. Furthermore, India’s life expectancy is expected to hit 82 in the year 2100 as per United Nations estimates as against 70.19 in 2022. As a consequence of the increasing life expectancy, which can be attributed to improved healthcare systems and healthier lifestyles, the number of people requiring old age income is bound to go up. Before this fiscal year’s budget was announced, NGOs and eminent economists had demanded an increase in the budget allocation to pensions. The increasing gap between the older and the younger generation and the uncertainty of old age were the key reasons for raising such a demand.

Furthermore, budgetary allocation to pensions has gone up each fiscal year. India’s pension budget has been ballooning, with 2.3 lakh crore being allocated to pensions, higher than that assigned to the healthcare, education, and energy sectors. Even in the defence sector, budget allocation to pensions has jumped 15.5% in FY 2023-24.

This is after India had already launched the National Pension Scheme under which individuals can save and invest their pension contributions in pension fund schemes.

Furthermore, the government increased the retirement age from 58 to 60 in 1998 on the recommendation of the 5th Pay Commission. This rising burden on the government exchequer has led to demands for increasing the retirement age in India.

However, the issue cannot be resolved by increasing the retirement age, considering the increasing workforce in India and insufficient job opportunities. Additionally, reports reveal that India’s employable talent has risen in recent years and women’s participation in the workforce has registered an increase. In that case, increasing the retirement age would restrict the employment of the upcoming younger workforce. Thus, in the Indian context, the issue of rising pension bills has to be dealt with through other pragmatic measures.

A Balancing Act

Given the relatively young population and increasing workforce in India, it might neither be necessary nor feasible to raise the retirement age. However, the issue of rising pension bills (a transfer payment) impacting capital expenditure in sectors like defence requires creative solutions. One such solution could be to explore alternate sources for financing pensions like increased employers’ contribution to pension funds and widening of the direct tax base. The former can be used to substitute a portion of government expenditure on pensions and make such systems more sustainable in the long run. The latter, by augmenting the government’s revenue, can enhance the ability to pay for pension schemes without limiting the resources available for undertaking development activities. India could also borrow from the Singaporean model which does not permit employers to dismiss employees on the grounds of age. Retirement at times leads to a sense of loss of identity. The opportunity for voluntary re-employment after retirement till the age of  70 on flexible terms decided by both the employer and the employee in Singapore gives a sense of security and autonomy to the workers. Owing to decades of work experience, they could engage in teaching, mentoring, and consultancy services so that their expertise, technical knowledge and skills can be put to valuable use. This will help the younger generation gain from the valuable experience of the seniors thereby promoting multigenerational bonding as envisaged in the National Policy for Senior Citizens.

Thus, increasing the retirement age to deal with the issue of rising pension bills in India is a half-baked solution. The issue needs to be combated using alternatives like exploring alternative sources for financing pension bills and providing opportunities for voluntary re-employment post retirement. 

The authors are undergraduate students of law at the National Law School of India University, Bengaluru. Kanishk is an editor at LSPR while Oorja is an observer.

Did I write this or did ChatGPT?

Prem Parwani

In December 2022, the rollout of the Artificial Intelligence (‘AI’) software ChatGPT immediately made headlines. Its ability to respond to human prompts with precision and depth skyrocketed the site to popularity.
ChatGPT has the potential to be a game-changer for academic writing given that it can respond to novel and complex prompts in a matter of seconds. Similar prompts can take a human a substantial amount of time to research, process and then articulate. For example, it can write a five-hundred-word essay comparing the works of Nietzsche, Kant and Freud, or apply legal principles and case law to any fact scenario entered by the user. It can be used not just as a research tool, but also as a problem solver.
Naturally, this has raised concerns of plagiarism in the world of academia, with some universities even banning its usage. How do we check the plagiarism of content through ChatGPT? Is using a novel idea created by ChatGPT plagiarism? In this article, I will analyse these plagiarism-related challenges for academic writing, and how academia can address them.

Detecting Plagiarism

Broadly, academic writing will have to contend with two forms of plagiarism – content-plagiarism and idea-plagiarism.1 Content-plagiarism refers to merely copying parts of the text generated by the AI, while idea-plagiarism refers to the substantive bits; using the ideas, concepts and analyses that the AI provides. This distinction is relevant because methods of detecting plagiarism differ in each case.
Detecting content-plagiarism is easier than detecting idea-plagiarism. Software such as ‘GPTZero’ has been designed to detect whether or not text is written by AI. This software detects such text by measuring their ‘perplexity and burstiness’, which is a measure of how complicated, random and structurally uneven the text is. Human text tends to be more random and uneven, while text generated by AI is usually simple and evenly structured. Using this metric, GPTZero has a 98% success rate in detecting whether a text was written by a human or an AI. Further, in a study where experienced academicians were asked to distinguish whether or not a text was written by ChatGPT, the academicians judged correctly 68% of the time. University professors have also testified that AI has a visibly “vague and formulaic” way of writing, which professors can spot with ease.
However, detecting content-plagiarism through these anti-plagiarism tools is not entirely straightforward. First, it is possible to artificially ‘humanise’ this text. Asking ChatGPT to ‘make the text more human’ makes it harder for GPTZero to detect it as AI-generated. Further, the text can be run through paraphrasing software and manually edited to increase its ‘perplexity and burstiness’2; i.e., make it seem more human. Second, the ability to detect

plagiarism also depends on the nature of the text. Creative text such as hypothetical scenarios and song compositions generated by ChatGPT remain almost entirely undetected by GPTZero.3 Third, this software might turn up errors by recognising human text as written by AI. Since software like GPTZero measures the ‘perplexity and burstiness’ of a text, human text which is well-written and structured can be wrongly recognised as AI text.4 Thus, although tools exist to detect content-plagiarism, they are far from perfect.
But how would idea-plagiarism be detected? This issue is slightly more complex since it requires us to consider how ChatGPT constructs its responses. Essentially, it uses the repository of publicly accessible data (which is just text-based data) as raw material to construct its responses. It then applies ‘algorithmic machine learning’ to this data, which means that it analyses the pattern and structure of this data. It ‘learns’ from this data and is thus able to apply patterns and structures across new data. Put simply, it is a brain that educates itself using data. This enables ChatGPT to create entirely novel ideas.
In response to being asked whether it can produce “novel analyses and ideas”, ChatGPT responded that it can.5 It can even be specifically asked to generate novel responses. Given that we cannot access the sources from which ChatGPT constructs its novel ideas, it is difficult to know when someone has come up with a concept of their own, or merely used one generated by ChatGPT. The issue here is that we have no straightforward way of knowing where the idea comes from. Is this the AI’s analysis? Is it relying on existing ideas? If ChatGPT’s responses and ideas resemble those in existing literature, then a simple literature review would be sufficient to check for idea-plagiarism. But for novel responses, detecting idea plagiarism seems to be difficult for now.

But is this plagiarism?

Let’s take a step back. If a student uses a novel idea created by AI, is it even plagiarised? Plagiarism has been traditionally defined as “passing off someone else’s work as your own.” But when the user inputs a prompt and receives a response, whose work is it– the AI or the user?
There are two different ways of viewing this issue. First, that the idea is produced by the user himself, since it is the user who generated the response by entering a specific prompt. Since the novel idea would not have been generated if not for the prompt, it can be argued that the AI’s response is a product of the user’s work. Further, it can also be argued that the definition of plagiarism does not accommodate AI, as AI cannot be ‘someone else’. Thus, ChatGPT can be seen as a form of research tool with the user being the primary producer of the idea.

Second, that the idea is produced by ChatGPT and not the user. It can be argued that the degree to which the user was involved in generating the final response is minimal as compared to the AI; i.e., the user plays an insignificant role in the AI’s final output. Although the user entered the prompt, the AI compiled and articulated the idea in question. Thus, ChatGPT would be the originator of the idea. It is clear from this issue that the emergence of AI has brought the need to reshape the definition of plagiarism itself. Multiple considerations must be balanced to shape plagiarism in the context of AI. For example, the goal to encourage independent thinking must be balanced with the utility and efficiency that AI tools offer. To do this, academia must keep pace with technology.

What can academia do?

Before proceeding to detect plagiarism, there is a need for discourse in academia about what plagiarism is and to what extent AI stretches its definition. Once this discourse has developed, academia can focus on detecting content and idea plagiarism. To do this, it has to keep up with the evolution of AI. This involves understanding and harnessing the technology that goes behind AI’s creation of responses. It has even been suggested that AI tools be developed specifically to detect plagiarism by other AI. As plagiarism evolves, so must our equipment to deal with it.
For starters, content plagiarism can be dealt with by creating awareness of the writing style of AI, and idea plagiarism requires thorough literature reviews. These plagiarism checks may be supplemented by AI software designed to detect AI plagiarism. Such software is bound to develop, given that even the creators of ChatGPT have also expressed interest in checking plagiarism. Regardless of the methods adopted, academia must foray into AI if plagiarism is to be checked.


In this article, I divided forms of plagiarism by AI into idea and content plagiarism and analysed the methods and hurdles in detecting them. Further, I explored the broader question of whether idea plagiarism from AI can be called ‘plagiarism’. Lastly, I highlighted the need for academia to evolve alongside AI, so that both content and idea plagiarism can be checked.
ChatGPT is merely the beginning of human-like AI, as Google has announced that it will shortly release its own “ChatGPT killer” named Sparrow. As AI evolves, it is to be seen to what extent its capabilities can affect academia, and what measures will have to be taken to check plagiarism.

[1] It is important to note here that these two forms of plagiarism are not mutually exclusive. This distinction is made because the methods applied to detect them will differ.
[2] AI models construct their responses by analysing existing data and ‘predicting’ what the next word would be in a sentence. Thus, GPTZero is also a form of AI that predicts what the next word might be and compares this with ChatGPT. If the comparison shows that the prediction matches, that means that the text entered is AI generated. If not, it is human generated and has more ‘perplexity and burstiness’.
[3] The author has found this pattern by entering multiple prompts on ChatGPT and GPTZero. The same may be verified by using the two software.
[4] The author has found this pattern by entering multiple prompts on ChatGPT and GPTZero. The same may be verified by using the two software.
[5] This answer was received in a response to asking ChatGPT whether it can construct novel analyses of its own.

The author is an undergraduate student at the National Law School of India University and an observer at LSPR.

Categories: Newsletter

Tagged as: