In early December, the Union Home Ministry placed a set of numbers before the Lok Sabha. In response to a question by Congress MP Shafi Parambil, Minister of State for Home Affairs Nityanand Rai disclosed state-wise figures of persons arrested and convicted under the Unlawful Activities (Prevention) Act (UAPA) between 2019 and 2023, as compiled by the National Crime Records Bureau (NCRB).

The totals are stark. In just five years, 10,440 people were arrested under UAPA across India. Over the same period, 335 were convicted. Even allowing for the time lag between arrest and trial, the pattern is difficult to ignore: UAPA continues to be invoked at scale, while the number of completed convictions so far remains relatively limited. In a statute where bail is difficult by design, this inevitably throws attention on what happens in the long space between arrest and judgment.

This is not an argument that UAPA should not exist. It is a question about how it is working in practice and whether its pre-trial structure is operating within the discipline that Article 21 requires.

The NCRB-based annexure to the Lok Sabha reply gives a year-wise national picture:

• 2019: 1,948 arrests; 34 convictions

• 2020: 1,321 arrests; 80 convictions

• 2021: 1,621 arrests; 62 convictions

• 2022: 2,636 arrests; 41 convictions

• 2023: 2,914 arrests; 118 convictions

Rough conviction ratios must be treated cautiously, because convictions in a given year may arise from cases filed earlier. But even with that caveat, the consistent gap between the number of people brought into the system and the number ultimately convicted is striking.

This sits within an already concerning longer-term pattern. Earlier data presented in Parliament for 2016–2019 showed that only a small percentage of UAPA cases ended in conviction. A 2022 study by the People’s Union for Civil Liberties (PUCL), drawing on NCRB data for 2015–2020, calculated that more than 8,371 people were arrested under UAPA while only 235 were convicted. Taken together, these figures confirm that a very large part of UAPA’s practical impact lies in the pre-trial phase.

The 2019–2023 dataset also reveals how unevenly UAPA is used across the Union.

• Jammu & Kashmir accounts for 3,662 arrests over five years — more than one-third of all UAPA arrests nationwide — but only 23 convictions.

• Uttar Pradesh follows with 2,805 arrests and 222 convictions, the highest number of convictions among all States. Yet, more than 9 out of 10 people arrested there have not resulted in conviction so far.

• Assam, Manipur and Jharkhand show high arrest figures with negligible convictions. For example, Assam recorded 755 arrests and only one conviction; Manipur 1,168 arrests and 1 conviction.

There is no straight line from these numbers to a conclusion about misuse. Conflict situations, preventive strategies and evidentiary challenges may all play a role. But in several regions, particularly Jammu & Kashmir and parts of the North-East, UAPA appears to operate less like an ordinary criminal law and more like a continuous exceptional regime under which large numbers of people can be arrested and held in custody while very few cases reach conviction.

Equally important is what the Lok Sabha was not told. The Home Ministry’s reply explicitly states that NCRB does not maintain state-wise data on how many people are currently in prison under UAPA. Parliament has been told how many have been arrested and how many have been convicted, but not how many are still inside, how long they have been in custody, or how many have been acquitted, discharged, or released on bail.

For a law whose most distinctive feature is its restrictive bail regime, this is not a minor omission. It makes it much harder to see how UAPA is working on the ground.

To understand why high arrest numbers and low conviction figures coexist, we have to look at how UAPA structures bail and investigation.

Section 43D(5) of the Act, as interpreted by the Supreme Court in National Investigation Agency v. Zahoor Ahmad Shah Watali (2019), requires courts to deny bail if, on a reading of the case diary and charge-sheet, the allegations appear “prima facie true”. The Court held that at the bail stage, judges must broadly accept the prosecution’s version and the admissibility of its material, instead of weighing reliability in any detailed way. In practice, once the investigating agency files a detailed charge-sheet, bail becomes exceptionally difficult.

This approach was partially tempered in Union of India v. KA Najeeb (2021), where the Supreme Court held that constitutional courts may grant bail despite Section 43D(5) where prolonged incarceration and delayed trial undermine Article 21. In other words, even under UAPA, there is a point at which time itself becomes a constitutional factor.

Recent decisions have applied this logic. Earlier this year, the Supreme Court granted bail to Shoma Kanti Sen, arrested in 2018 in the Bhima Koregaon case, after nearly six years in custody. The Court expressly took account of the long period of incarceration and the slow progress of trial.

And in a recent order, the Supreme Court sharply criticised a situation arising from Assam, A man arrested under UAPA had been in custody for over two years without a chargesheet. The Bench of Justices Vikram Nath and Sandeep Mehta described the situation as “appalling,” pointing out that even under Section 43D, investigation can only be extended by a maximum of 180 days.

These decisions do not dilute the seriousness of UAPA offences. They show the Court trying to ensure that a statute designed for exceptional threats does not result in open-ended pre-trial deprivation of liberty.

The official figures are numerical; their consequences are lived in individual cases.

The world came to know the name of Father Stan Swamy, an 84-year-old Jesuit priest and tribal rights defender arrested in October 2020 under the UAPA in the Bhima Koregaon case. He suffered from Parkinson’s disease, was repeatedly denied bail and died in judicial custody in July 2021.

In the same cluster of prosecutions, former Delhi University faculty member Hany Babu was arrested in July 2020 and remained in custody for more than five years before the Bombay High Court granted him bail in December 2025, noting that prolonged incarceration without any near prospect of trial completion raised serious concerns under Article 21.

Scholar Anand Teltumbde spent 31 months in Taloja Central Prison as an undertrial before receiving bail from the Bombay High Court in 2022. In his prison memoir The Cell and the Soul, he describes incarceration under a special statute as a suspension of ordinary life: work, family and dignity are all placed on hold while the case moves at a pace determined by the system.

More recently, Umar Khalid, a former JNU scholar and activist, has become an emblematic example of UAPA’s temporal logic. Arrested in September 2020 in connection with the Delhi violence conspiracy case, he has now spent more than five years in Tihar Jail without his trial concluding. His bail pleas have been repeatedly denied; only now is the Supreme Court hearing his appeal.

How a person experiences UAPA also depends on whether they can effectively defend themselves. UAPA trials typically involve thousands of pages of electronic records, forensic material and witness statements. Preparing a defence demands sustained time, expertise and continuity. For families with resources, this is difficult but possible. For many others, it is simply out of reach.

Nationally, studies using NCRB data have found that only a small minority of undertrial prisoners report using free legal aid. Projects like the Fair Trial Programme in Maharashtra have documented that undertrials often meet their legal aid lawyers infrequently, that continuity across hearings is weak and that poor coordination between prisons, courts and legal-services authorities undermines effective defence.

When such structural weaknesses meet a law like UAPA, the inequalities deepen. An accused person with private counsel may be able to insist on timely investigation, press for default bail where chargesheets are delayed, or seek relief under Najeeb when custody becomes excessive. An indigent accused, reliant on an overburdened legal-aid system, may remain in custody simply because the system cannot keep pace with the demands of the case.

This does not negate the need for a counter-terror law. It does mean that defence capacity and legal aid are not peripheral questions in UAPA cases; they are central to whether the statute is operating in a constitutionally fair manner.

The new UAPA figures from Parliament have done something important: they have turned long-standing concerns into official data. We now know that:

• arrests under UAPA are rising,

• completed convictions so far remain relatively few

• the State does not maintain systematic, state-wise data on undertrials and custody duration.

These facts do not, by themselves, prove misuse. But they do create a duty of closer scrutiny.

A serious response need not be abolitionist. It can begin with a few clear commitments:

Transparency: Publishing annual state-wise data not only on arrests and convictions, but also on undertrials, duration of custody, bail outcomes, acquittals, discharges and withdrawals.

Time-bound investigation and trial: Ensuring that the 90–180 day investigation limits are respected in practice and treating violations as serious, as the Supreme Court did in the Assam case.

Article 21–sensitive bail: Applying the principles of Najeeb, the Shoma Sen order and similar cases so that prolonged incarceration does not become a default outcome in special-law prosecutions.

Strengthening defence: Investing in specialised legal aid capacity for UAPA and other special statutes, with adequate remuneration and manageable caseloads.

The latest UAPA data does not tell us who is guilty and who is innocent. That remains a matter for courts to decide in individual cases. What the data does show is that large numbers of people are being arrested under UAPA, relatively few cases have ended in conviction so far, and we do not yet have a full picture of how many remain in custody or for how long.

In a constitutional democracy, those facts are not neutral. They require that the strongest criminal laws be accompanied by the strongest safeguards - not only on paper, but in how bail is decided, how quickly investigations move and how well the poorest accused are represented.

The task is not to weaken UAPA’s ability to address genuine threats, but to ensure that its operation remains recognisably faithful to the guarantees under Article 21. The parliamentary numbers have started that conversation. It is now for Parliament, the executive, the Bar and the courts to decide whether they are willing to read those numbers as a signal for closer constitutional attention and to act accordingly.

Sahil Hussain Choudhury is a lawyer and constitutional law researcher based in New Delhi.