1 In 2019, Parliament introduced amendments to the Penal Code (Cap 224, 2008 Rev Ed) (the “Penal Code”) to specifically deal with the entire ecosystem of child abuse material: from the production to the distribution and consumption of such material. Among the provisions was s 377BK of the Penal Code (as introduced by s 120 of the Criminal Law Reform Act 2019 (Act 15 of 2019)) which makes it an offence to, among other things, possess child abuse material. The Minister for Home Affairs, in explaining the reforms to the Penal Code to target child abuse material, spoke on the “[t]errible harm … caused to these children who are used in the production of such material” (Singapore Parl Debates; Vol 94, Sitting No 103; [6 May 2019] (K Shanmugam, Minister for Home Affairs)). The Minister emphasised that “[a]part from sexual abuse, some children are physically abused; they are tortured as well.” The Minister also noted that the production of such material did not happen much in Singapore, but the reforms would have some extra-territorial effects, and had the effect of “criminalis[ing] the spectrum of offences to deal with every person involved, from consumers to the producers of such material.”

2 This appeal concerned an offender, the respondent, who was a consumer of child abuse material. The respondent was charged with possessing child abuse material, namely two still images and six videos, under s 377BK(1) punishable under s 377BK(2) of the Penal Code. Two charges ((a) a similar charge for accessing other child abuse material; and (b) a further charge under s 30(1) of the Films Act (Cap 107, 1998 Rev Ed) (the “Films Act”) for possessing 119 obscene films) were taken into consideration for sentencing.

3 Given that s 377BK of the Penal Code was a new offence that came into operation in January 2020, there was a dearth of reported sentencing precedents. This appeal thus provided us with an opportunity to consider the appropriate sentencing framework for an offence under s 377BK(1) punishable under s 377BK(2) of the Penal Code.

4 We appointed Mr Benny Santoso (“Mr Santoso”) as young independent counsel (the “YIC”) to assist us with the following questions:

5 The framework that we set out below would apply to possession cases in s 377BK(1) of the Penal Code (the “Possession Offence”) as this case involved possession of child abuse material, and not gaining access to child abuse material. As noted in Public Prosecutor v GED and other appeals [2023] 3 SLR 1221 (“GED”) at [41], it will not generally be appropriate for an appellate court to lay down a sentencing framework for an offence that is not before the court, for reasons of principle and practicality. It is not the role of the court – being a judicial rather than legislative or quasi-legislative body – to lay down sentencing frameworks for offences that are not before it. As a matter of practicality, any submissions on offences not before the court would be hypothetical and neither relevant to nor necessary for the disposal of the case at hand. Without the facts of an actual case before the court, it would be difficult to anticipate how various considerations may or should feature in the court’s approach to sentencing, or what significance should be accorded to those considerations, in that context. We therefore only provide guidance as to the appropriate sentencing framework that should be applied for the Possession Offence under s 377BK(1) punishable under s 377BK(2) of the Penal Code. We leave the sentencing framework for the offence of gaining access to child abuse material to be decided in an appropriate future case where it arises squarely for determination.

6 In the court below, the respondent pleaded guilty to the following charge, and consented to having the following additional two charges taken into consideration (“TIC”) for sentencing:

7 The respondent first grew curious about child pornography in 2020. On 7 May 2020, he used the TOR Browser (a web browser that anonymised its users’ web traffic) and found a search engine from which he could access websites on the dark web containing child abuse material. He paid Bitcoin for full access to one such website but failed to gain the log-in details. This incident was reflected in the TIC charge DAC-903560-2022.

8 Around June 2021, the respondent joined a Telegram chat group where pornography (including child abuse material) was shared. He knew that members of the chat group shared child abuse material as he had downloaded some of the child abuse material. This downloaded material formed the basis for the proceeded charge DAC-903561-2022. There were two still images showing fully nude girls (who appeared to be below 14 years of age) standing next to one another. There were six videos between 51 seconds long (shortest) and 37 minutes and 46 seconds long (longest). The videos showed young girls engaging in various sexual acts. As an example, one of the videos showed: (a) a girl (who appeared to be below nine years of age) touching her vagina with her legs apart, (b) a man performing cunnilingus on her; and (c) the man then penetrating her vagina with his penis, with the girl appearing to be in pain from the penetration.

9 The District Judge (“DJ”) sentenced the respondent to six weeks’ imprisonment in respect of the single proceeded charge DAC-903561-2022. The DJ’s reasons are found in Public Prosecutor v Randy Rosigit [2023] SGDC 59 (the “GD”).

10 The DJ considered that the relevant sentencing factors could be grouped into two heads: “possession” and “child abuse material” (GD at [43]).

11 “Possession” was focused on the offender’s level of involvement with the child abuse material and his motives, assessed based on the manner in which the s 377BK offence was committed (GD at [44]). It accounted for the following non-exhaustive factors (GD at [45]–[58]):

(a) method of gaining possession, which could be sub-divided into the following sub-factors:

(i) degree of planning and preparation;

(ii) level of sophistication;

(iii) anonymity of the offender;

(b) length of possession, which could be sub-divided into the following sub-factors:

(i) duration of offending behaviour; and

(ii) persistence in offending behaviour.

(c) type of possession: how and in what form the child abuse material was stored; and

(d) motive for possession: personal use, financial gain or other illicit purposes (eg, sharing with others, promoting child abuse material, grooming).

12 “Child abuse material” required examination of the following factors (GD at [61]–[72]):

(a) nature of images and/or videos, which could be sub-divided into the following sub-factors:

(i) parts of body revealed (degree of vividness and intrusiveness);

(ii) age and number of victims; and

(iii) acts depicted;

(b) volume of child abuse material involved; and

(c) nature and extent of distribution, if any.

13 The DJ held that the length of recordings was not relevant. At most, video length might be relevant to progression or persistence in offending (GD at [70]–[71]).

14 In applying the above sentencing factors to the facts, the DJ made these findings:

(a) Method of gaining possession: The respondent’s method of offending was simple and straightforward – child abuse material was downloaded from a Telegram group, without sophisticated searches or browsers, special access or permission. The materials were stored in his personal devices, without him being surreptitious or using unlawful means. There was no attempt made to mask his identity or to avoid detection (GD at [45]–[48] and [99]).

(b) Length of possession: The respondent’s progression in offending was slight and at a low level. Four months was spent on a single Telegram group, and the downloading took place over a short span of two months. There was some persistence in accessing and obtaining child abuse material: first in accessing them on the dark web, then downloading just one video at a time from Telegram, before downloading four at one go. Downloading ceased because his home was raided, and not because of any insight into his conduct. While there was persistence, this was sporadic (GD at [49]–[53] and [99]–[100]).

(c) Type of possession: The child abuse material was stored on the respondent’s mobile phone, without steps taken to avoid detection or hide the child abuse material. His possession of child abuse material was quickly discovered during a police raid. There was no evidence of copying or duplication onto multiple devices (GD at [54]).

(d) Motive for possession: The respondent possessed child abuse material solely for personal use (GD at [55]–[58] and [99]).

(e) Nature of images: The nature of the child abuse material fell within the lower end of the spectrum of cases involving possession of child abuse material. There was a high degree of exploitation – the child abuse material was graphic, with nudity or exposed genitalia. Some of the girls appeared pre-pubescent, with the oldest appearing to be no more than 11 years old. One girl appeared to be in pain from penile-vaginal sex. But there was no torture, gratuitous cruelty, violent or deviant sexual practices, or pain resulting from these (GD at [61]–[68] and [98]).

(f) Quantity of child abuse material involved: There was a low volume of child abuse material, with two images and six videos. Precedent cases saw higher volume of child abuse material being obtained over an equally short amount of time (GD at [69] and [98]).

(g) Nature and extent of distribution: There was no evidence of distribution (GD at [72]).

15 As for the two TIC charges, the DJ held that they only merited a moderate uplift in sentence. For the s 377BK access charge, while the respondent took steps to avoid detection when downloading the TOR browser to access the dark web, which increased his level of offending, there was no evidence of him having paid for any of the materials that he finally gained access to, such that he had advanced the profit incentive of others in the supply chain. Furthermore, while the access charge demonstrated the respondent’s singular commitment to gaining access to child abuse material, this was not a continuous pattern of offending, but two separate occasions of first accessing child abuse material on the dark web, and then downloading child abuse material through the Telegram group. As for the Films Act charge, the volume of obscene films was not small, but also not so large either – especially when compared to precedent cases (GD at [73]–[76] and [101]–[102]).

16 The DJ concluded that the case fell “close to the least aggravated end of the spectrum of cases” (GD at [103]). The precedent cases showed that highly aggravated cases attracted five months’ imprisonment, while four weeks’ imprisonment applied for less egregious offending. On the facts, a sentence of less than four weeks’ imprisonment would not accurately reflect the two aggravating factors in the case: the content’s explicit and intrusive nature, and the respondent’s persistence in obtaining them (GD at [103]). Considering the low degree of offending, the lack of other compelling aggravating factors, and the mitigating factors (that the respondent pleaded guilty, had no antecedents, and did not reoffend after his home was raided), the DJ concluded that six weeks’ imprisonment was appropriate (GD at [77], [106]–[107]).

17 The Prosecution, being dissatisfied with the sentence imposed by the DJ, appealed against the sentence. There was no cross-appeal by the respondent.

18 On appeal, the Prosecution submitted that the respondent’s sentence should be increased to at least six months’ imprisonment. This proceeded from a starting sentence of 12 to 15 months’ imprisonment, which was justified by the “inherent severity” of child abuse material offences and the type of child abuse material possessed by the respondent. The reduction of this starting sentence to six months’ imprisonment balanced the respondent’s guilty plea and lack of antecedents on the one hand, with the TIC charge of accessing child abuse material.

19 The Prosecution submitted that the sentencing framework employed for s 377BK offences must “embody a strongly deterrent imperative”, bearing in mind that offences relating to child abuse material were “uniformly and profoundly abhorrent”. At the same time, the framework must be sensitive to the realities and constraints faced by investigators and criminal law practitioners. The framework should give effect to these principal sentencing factors:

(a) the quantity of child abuse material possessed;

(b) the nature of child abuse depicted;

(c) whether the offender specifically sought out especially gratuitous child abuse material (eg, those depicting torture, cruelty, bestiality, very young children, or children who were drugged or intoxicated);

(d) whether the offender took elaborate steps to evade detection;

(e) whether the offender intended or actually distributed or traded child abuse material, particularly when done for profit; and

(f) whether the offender showed persistence in obtaining child abuse material (typically evidenced by a long offending period, being a committed member of a community dedicated to child abuse material, and repeated interactions with the child abuse material possessed).

20 While the Prosecution considered both a multiple starting points approach and a two-stage, five-step sentencing framework set out in Logachev Vladislav v Public Prosecutor [2018] 4 SLR 609 (“Logachev”) to be capable of accommodating these features, on balance, it preferred the former. The Prosecution’s framework used the quantity of child abuse material as the principal sentencing anchor, with the following starting sentences applicable to offenders who claimed trial:

21 Within the applicable sentencing range, the type of child abuse material would then determine the indicative starting sentence. The court should then consider if any of the other principal sentencing factors (at [19] above) applied, such as the child’s age or the presence of additional acts of torture or cruelty. The presence of any such factor should warrant a “significant uplift”, possibly even by one or two sentencing bands.

22 The Prosecution submitted that its framework was less granular than Mr Santoso’s proposed Logachev-style framework and would avoid clustering sentences at the very lowest end of the sentencing range.

23 The Prosecution submitted that the respondent’s sentence of six weeks’ imprisonment was inadequate. Applying the Prosecution’s proposed sentencing framework, the starting sentence should have been about 12 to 15 months’ imprisonment. They further submitted that the similar offence taken into consideration was aggravating, and it was especially aggravating that the respondent had searched and paid for child abuse material on the dark web more than a year before committing the present offence. After taking this aggravating factor into account and calibrating the starting sentence downwards to account for the respondent’s plea of guilt, a sentence of about eight to ten months’ imprisonment would have been appropriate. For the purpose of the present appeal, the Prosecution sought a sentence of at least six months’ imprisonment.

24 Mr Santoso proposed a Logachev-style framework for the following reasons:

(a) The court’s consideration of harm and culpability factors should not be separated into different stages of the sentencing process, as consideration of harm and culpability would likely overlap (eg, the quantity of child abuse material possessed might be an aspect of both the harm caused and the offender’s persistence in offending).

(b) Equal weightage should be given to both harm and culpability factors, as they were “equally weighty factors”. This mitigated the difficulty arising from the potentially limited utility of the harm factors available to the sentencing court, given that the sentencing court was unlikely to have a clear picture of the subjective aspects of harm. This was because the precise identity of the victim of a s 377BK offence would often not be known to the court.

(c) The sentencing process was more robust when the court considered two axes in reaching the indicative starting point. This was in contrast to the two-step sentencing bands approach (see, eg, Ng Kean Meng Terence v Public Prosecutor [2017] 2 SLR 449 at [39]), which only presented a single axis on which the sentencing bands resided. Relatedly, a three-by-three matrix, such as the Logachev framework, offered greater methodical consistency, by providing five indicative starting point sentencing ranges as opposed to three indicative ranges under the two-step sentencing bands approach.

(d) The benchmark and single starting point approaches were unworkable as the factual circumstances in s 377BK offences did not present themselves in only one or two archetypal ways.

25 We will examine in further detail below Mr Santoso’s proposed framework when we set out our guidance on the appropriate framework to adopt for the present offence.

26 The Defence agreed with Mr Santoso’s submission that a Logachev-style framework was the most appropriate sentencing framework for the s 377BK offence. The Defence submitted that the respondent’s offence was one of slight harm and low culpability and argued that the DJ’s sentence of six weeks’ imprisonment was in line with the sentence arrived at after the application of Mr Santoso’s proposed framework.

27 The Defence submitted that the harm and culpability for the respondent’s offence should be pegged at the lowest end of the spectrum and a starting point of an imprisonment term not exceeding three months was appropriate. The Defence noted that only two out of eight items of child abuse material possessed by the respondent depicted penile penetration. There was also no evidence of multiple non-victims or sexual abusers present in the child abuse material, and none of the child abuse material involved children of a significantly young age. The Defence submitted that the respondent was a “one-off offender who sourced for [child abuse material] out of mere curiosity and sheer boredom during the COVID-19 pandemic” [emphasis in original] and was thus of a lower level of culpability than a sex addict or a paedophile. The Defence pegged the respondent’s level of planning, premeditation and sophistication at the lowest end of the spectrum and argued that this calibration was supported by the fact that he was quickly arrested after offending. The Defence submitted that the respondent’s role was limited to being that of a mere member of a Telegram group with no evidence that he had shared child abuse material either within the Telegram group or outside of the group.

28 In considering the offender-specific factors to make adjustments to the starting point, the Defence submitted that:

(a) the respondent’s TIC charges should carry limited aggravating weight “given that they [bore] close connection to the proceeded charge and … the relevant sentencing factors [had] sufficiently been factored into arriv[ing] at the starting point sentence for the proceeded charge”;

(b) he pleaded guilty early and co-operated with authorities throughout the course of investigations;

(c) he was genuinely remorseful and regretted his actions;

(d) he did not reoffend while on bail; and

(e) he was completely untraced.

29 Taking these offender-specific factors into account, the defence submitted that there should be a downward calibration of the proposed starting point sentence of three months’ imprisonment to six weeks’ imprisonment.

30 This appeal raised two broad issues for our determination.

(a) First, what was the appropriate sentencing framework for the Possession Offence under s 377BK(1) punishable under s 377BK(2) of the Penal Code?

(b) Second, applying that framework, what was the appropriate sentence for the respondent?

31 We begin by setting out the text of s 377BK of the Penal Code in full:

32 Before we set out the appropriate sentencing framework for this offence, we make two preliminary points concerning the animating principles in formulating a sentencing framework, and its appropriate scope.

33 We must emphasise – again – a point that various appellate courts setting out sentencing frameworks have stated repeatedly. Sentencing frameworks should provide workable guidance to guide sentencing courts towards an appropriate sentence in each case using a methodology that is broadly consistent: Mohd Akebal s/o Ghulam Jilani v Public Prosecutor and another appeal [2020] 1 SLR 266 at [20(b)]. Sentencing frameworks should not aim for mathematical precision because the exercise of sentencing is largely a matter of judicial discretion and requires a balanced judgment and assessment of myriad considerations: Angliss Singapore Pte Ltd v Public Prosecutor [2006] 4 SLR(R) 653 at [13]. Sentencing frameworks should therefore be constructed with a view on clarity and with a focus on avoiding excessive complexity and consequent unworkability: Goh Ngak Eng v Public Prosecutor [2023] 4 SLR 1385 at [43]–[44]. They need not cater for all eventualities that might arise: Leong Sow Hon v Public Prosecutor [2021] 3 SLR 1199 at [53(c)]. In essence, sentencing is not a fine-grained analytical exercise but a broader inquiry that is designed to provide structure, clarity and consistency. Sentencing frameworks need to be understandable and easily applicable. Thus, we are of the view that some of the suggested distinctions and inquiries proposed by the Prosecution and the Defence are impractical to incorporate in the sentencing process.

34 We note that the Prosecution submitted that this court should endorse, for the purpose of cases where the police has seized offending material exceeding 100 items from an accused person, a system where the police would conduct a random sampling of up to 100 items of material in an offender’s possession, without examining or detailing all the material. We decline to decide on this issue as it is not necessary to deal with this question of sampling in the present case, which involved a small volume of child abuse material, all of which were examined and detailed in the statement of facts. We further note that the issue of sampling may go to the contents of the appropriate charge to frame against the accused person, and this is not an issue that this court, in hearing an appeal against the sentence imposed arising from a plea of guilt, should decide on.

35 We are of the view that the multiple starting points approach preferred by the Prosecution is unsuitable for the present offence. It is overly-blunt, in that it focuses overwhelmingly on one metric: the quantity of child abuse material involved. This approach fails to differentiate between salient factors in a child abuse material-related offence, such as the difference between still images and videos, and the difference in the contents of the material. We will later on set out the various harm and culpability factors which are relevant in sentencing for the present offence. It suffices to say here that when one focuses only on the quantitative metric of how many items of child abuse material are present in a given case, one misses all of the nuances in the harm and culpability factors that help inform the true criminality and gravity of the case. In addition, the Prosecution’s suggested multiple starting points approach tends to provide an anchoring effect on the volume of child abuse material involved, which may, again, distort the assessment of the true gravity of the offence.

36 We agree largely with Mr Santoso’s formulation of a Logachev-style framework for the Possession Offence. A Logachev-style framework considers harm and culpability equally. This is important for the Possession Offence because harm and culpability are the two principal factors that would drive the level of sentence that is appropriate in order to deter such conduct. We turn to explain each of the components of the Logachev-style framework for the Possession Offence.

37 At the first step of the Logachev-style framework, the court will have regard to the relevant offence-specific factors and identify: (a) the level of harm caused by the offence; and (b) the level of the offender’s culpability (Logachev at [76]). We first examine the offence-specific factors going towards harm in greater detail.

38 We emphasise that harm is a very serious factor in the present offence. Offences involving child abuse material, and related offences such as child sex tourism, are a particularly egregious variety of offending because of the extreme vulnerability of the victims. The direct harm to the children involved in these cases is plain to see. It would be artificial to suggest that the real harm is only the long-term psychological effects on the child victims. That there are no victim impact statements is also beside the point – we agree with Mr Santoso’s observation that in many cases involving possession of child abuse material, the child victims involved in the production of the child abuse material may not have been precisely identified or located, much less brought before the court.

39 To deal with this harm, it is important to focus on the demand side for such material. The Penal Code Review Committee, which recommended the creation of new offences relating to child abuse material (Penal Code Review Committee, Penal Code Review Committee Report (August 2018) (the “PCRC Report”) at pp 126–132), flagged that technology had facilitated both the distribution of and the demand for child abuse material (PCRC Report at p 125, para 7). The PCRC Report further noted (at p 129, para 34) that the relationship between the maker of child abuse material and the consumer can be likened to the relationship between receivers and thieves. Without a market for stolen goods, there would be no incentive to steal. Similarly, if not for the consumers, there would be no market for those who abuse children by creating this material. Much of the material is not created from within Singapore even if the consumers of such material may be based in Singapore, and this was indeed recognised by Parliament (see [1] above). It is especially important for courts in the countries where there is demand for child abuse material to come down hard on such consumers because the legal frameworks in the supply side countries may not be as robust or reliable.

40 It is true that there are separate offences for different types of activity – such as producing child abuse material (s 377BH of the Penal Code), distributing or selling child abuse material (s 377BI of the Penal Code), and advertising or seeking child abuse material (s 377BJ of the Penal Code) – but this does not change the fact that there is real harm to the victims whose abuse form the content of child abuse material, who are losing their childhood, their innocence and ultimately their human dignity. The harm results largely because there is demand for such child abuse material. The consumer of child abuse material may not be the one producing the material, but he would know that because he is a consumer, he creates the demand for such material and incentivises others to produce it for profit. We thus agree with Mr Santoso’s identification of market-making harm as one of the important overarching harm factors to be considered in sentencing for the Possession Offence.

41 We also agree with Mr Santoso on most of the relevant harm factors that he had identified.

42 Concerning the type and nature of the acts depicted in the child abuse material, while the sequencing of the acts in the definition of child abuse material in s 377C of the Penal Code is not necessarily determinative in terms of Parliament’s intention, it is helpful as a matter of common sense to note that, in general, that sequencing does reflect a sensible way to approach the gravity of the acts, and the direct harm that the acts cause to the child victim involved in the production of the child abuse material. We reproduce the relevant part of s 377C of the Penal Code for reference:

43 The direct harm that the acts cause to the child victim involved in the production of the child abuse material refers to both the physical harm that is caused in cases of torture or sadistic acts, the psychological harm which can be presumed in most cases, and the harm from the loss of dignity, innocence and privacy. We do not see any value in analysing these harms (physical, psychological, moral, etc) as though they are distinct categories of harm or that they are of relatively different gravity. In addition, we observed at the hearing that whilst s 377C(1) of the Penal Code specifies, as part of the sequencing of acts, that “child abuse material” includes material that depicts an image of a child “as a victim of sexual abuse”, this particular component of the sequence is not particularly useful in a sequencing of the type and nature of acts depicted in child abuse material in increasing level of harm. This is because all the child victims depicted in child abuse material are victims of sexual abuse in a sense, so this category might be over-inclusive if seen as a standalone category.

44 We thus modify Mr Santoso’s suggested scale of acts depicted in the child abuse material as follows:

45 We note that the Defence had suggested modifying Mr Santoso’s suggested scale of acts such that: (a) where there were multiple non-victims (ie, those performing sex acts on children) depicted in the child abuse material, this will contribute to a higher harm level than child abuse material without multiple non-victims, especially where there is a visibly significant age difference between the child and the non-victim; and (b) “there should be a distinguishment between oral penetration and penile penetration with penile penetration being a more egregious form of sexual abuse”.

46 We decline to overly-granulate the scale of acts by incorporating the Defence’s suggestions. These factors may well be considered by a sentencing court as relevant considerations in the court’s precise calibration of the harm entailed in the abuse suffered by the child depicted in the child abuse material. However, the key consideration, in setting out a sentencing framework, is whether Mr Santoso’s proposed scale of acts is sufficiently, but not overly, granular, and this consideration seeks to balance the competing considerations of providing sufficient guidance to first-instance courts to achieve consistency and transparency in sentencing, while preserving flexibility and easy-application of the framework so that first-instance courts can deal effectively with the myriad fact scenarios they are faced with. We observe that where appellate courts have set out Logachev-style sentencing frameworks for offences, the labels for the relevant offence-specific factors were typically set out at a high level of abstraction, and there was emphasis that sentencing frameworks should not be overly prescriptive. While sub-factors or considerations that go towards the broad offence-specific factors were discussed by the courts, these sub-factors and considerations were typically framed on the footing that they were relevant considerations that a first-instance sentencing court could take into account depending on the facts before the court: see, eg, Sue Chang v Public Prosecutor [2023] 3 SLR 440 at [45], [72] and [90], Logachev at [37]–[38] and [77], and Wong Meng Hang v Singapore Medical Council and other matters [2019] 3 SLR 526 at [32] and [39]–[41].

47 The Prosecution submitted that consumption of child abuse material might lead to addiction and escalated offending, with consumption being a gateway to the commission of other related offences such as child sex tourism (see, eg, Chan Chun Hong v Public Prosecutor [2016] 3 SLR 465 where the offender started with consuming child abuse material before his offending escalated into engaging in child sex tourism and arranging for others to do the same). We accept the Prosecution’s submission, to the extent that the submission put forth the proposition that consumption of child abuse material is pernicious, and a firm stance should be taken against it. However, we would caution that it is inappropriate to examine the potential for offenders to graduate to other more serious types of offences such as child sex tourism, as the law does not punish for intention, much less for predilection. To avoid doubt, to the extent that the facts before the court show that the offender had actually gained possession of child abuse material with escalating gratuity over time, that would be a legitimate factor to be considered for the purpose of sentencing.

48 We agree with Mr Santoso’s suggestion that the more exposure there is of private body parts of the victim child, the more aggravated the harm. Similarly, the number of different children depicted in each item of child abuse material is a relevant harm factor; the greater the number of different children depicted, the greater the harm. We also agree that the quantity of child abuse material is a relevant harm factor because it affects the extent of the demand that is being generated and the harm caused to the victims in the child abuse materials. The age of the child depicted in the child abuse material is another harm factor because it goes to the gravity of the impact on the victim, which can be inferred from the particular vulnerability of younger victims. If, however, the evidence shows that the offender deliberately sought out such child abuse material depicting very young children, then it would also go towards culpability because it shows an intention to target such vulnerable victims. The type of media is relevant only in the sense that a video is presumptively more injurious to the victim than a still. For the same reason, a much longer video is more harmful than a less lengthy one. The extent to which the victim can be identified in the child abuse material is another harm factor because it personalises the injury even more pointedly. In the rare case where there is available evidence of the subjective harms suffered by the victims of child abuse material, such as where a victim provides a victim impact statement, that should be considered as well. However, the absence of evidence of subjective harm cannot be taken to be a factor favouring the offender.

49 The harm factors canvassed above are non-exhaustive, and more factors may be identified as more cases come before the courts.

50 Turning to culpability, we agree that the following non-exhaustive culpability factors would warrant consideration in many instances of the Possession Offence:

(a) Planning, preparation, premeditation and sophistication. These are well-established aggravating factors: Logachev at [56]–[58].

(b) Attempt to conceal the offence. Attempts to conceal the offence have also been often regarded as a relevant aggravating factor: Vasentha d/o Joseph v Public Prosecutor [2015] 5 SLR 122 at [69].

(c) Group participation. To the extent that group participation (eg, active participation in a text messaging chat group where members share child abuse material) reveals insights into the offender’s attitude to this offence, that may be relevant to culpability. We further observe that group participation may also be a relevant consideration in respect of harm because group offending may encourage and spur more criminal activity and result in a higher degree of actual and potential harm (see, eg, Public Prosecutor v Ong Chee Heng [2017] 5 SLR 876 at [32]–[36]).

(d) Duration of the offending. It is well-established and commonsensical that an offence perpetrated over a sustained period of time will generally be more aggravated than a one-off offence: Logachev at [59].

(e) Offender’s motive. The offender’s motive in committing the offence is relevant: Ye Lin Myint v Public Prosecutor [2019] 5 SLR 1005 at [51]–[53]. Thus, for instance, where an offender possesses child abuse material with an intention of distributing it further, with an intention of making profit therefrom, or with an intention of using the material for blackmail or sexual grooming of others, these may be aggravating considerations.

51 The non-exhaustive harm and culpability factors canvassed above may be summarised as follows:

52 The harm caused by the offence may be categorised into one of three levels: slight, moderate or severe. The offender’s culpability may be categorised into one of three levels: low, medium or high.

53 The next steps are to identify the applicable indicative sentencing range, and thereafter to identify the appropriate starting point within that range: Logachev at [78]–[79].

54 Under s 377BK(2) of the Penal Code, an offender who has committed the Possession Offence shall be punished with imprisonment for a term which may extend to five years, and the offender shall also be liable to fine or to caning. The following sentencing matrix is appropriate for the Possession Offence:

55 Given that imprisonment is mandatory for this offence, none of the sentencing ranges in the sentencing matrix provide for a fine-only sentence. An imposition of a fine in addition to imprisonment may be warranted in certain situations, for instance, where there is a need to disgorge profits an offender made from his illegal behaviour: Public Prosecutor v Su Jiqing Joel [2021] 3 SLR 1232 at [37]. We leave open the possibility that there may be other situations where a fine in addition to imprisonment or imprisonment plus caning may be appropriate.

56 The identification of the appropriate cell of the sentencing matrix to situate each case will depend on the court’s assessment of harm and culpability using the factors considered at the first step above. Thereafter, the court will have to identify the appropriate starting point within that range, having regard once again to the level of harm caused by the offence and the level of the offender’s culpability: Logachev at [79].

57 At the fourth step of the Logachev-style framework, the court will make such adjustments to the starting point as may be necessary to take into account offender-specific aggravating and mitigating factors (Logachev at [80]). These offender-specific factors are generally applicable across all criminal offences and are therefore “well settled in our criminal jurisprudence” (GED at [110], citing Logachev at [63]). A non-exhaustive list of offender-specific factors would include:

58 Where an offender has been convicted of multiple charges, the fifth step in the Logachev-style framework is to consider the need to make further adjustments to take into account the totality principle. This step in the Logachev-style framework has been discussed extensively in other cases (see, eg, GED at [115]–[118]) and we do not propose to further discuss this step here, given that the respondent was only convicted of a single proceeded charge.

59 We applied the framework set out above to the facts before us.

60 We calibrated the respondent’s offence at the lower end of “moderate harm”, and at the high end of “low culpability”.

61 In relation to harm, the salient factors were:

(a) Quantity of child abuse material possessed: The net quantity of child abuse material possessed (two still images and six videos) was low.

(b) Type or nature of acts depicted: The respondent possessed multiple items of child abuse material that depicted acts that were at a high level of harm under the scale of acts (see [44] above). The videos showed, in some instances, sex acts including fellatio, cunnilingus, masturbation, mutual masturbation involving two young victims and penetration.

(c) Number of different children depicted: The still images involved ten fully nude girls who appeared less than 14 years old.

(d) Type of media: The respondent possessed not just still images, but videos as well, the two longest of which were 9 minutes and 41 seconds long and 37 minutes and 46 seconds long.

(e) Age of children: The ages of the victims (as they appeared to a reasonable observer) were as low as less than six years old, and generally less than 11 years old.

(f) Degree of identifiability of the child(ren): the victims appeared to be identifiable (in that frontal views of them were visible in the child abuse material).

62 The various harm factors fixed the present case at the lower end of “moderate harm”, chiefly because the fairly high level of harm displayed in the acts depicted in the child abuse material needed to be balanced against the low net quantity of child abuse material possessed by the respondent.

63 In relation to culpability, the salient factors were:

(a) Degree of planning, preparation, premeditation and sophistication:

(i) The respondent searched for, and found, child abuse material on the dark web using the TOR Browser. The TOR Browser was a platform that was designed to hide a user’s internet footprints. He also paid in Bitcoin for full access to a website with child abuse material but failed to gain the log-in details. Even though the respondent failed to get full access to the child abuse material website, he still gained some access to the child abuse material, as was clear from the TIC charge DAC-903560-2022. As pointed out by Mr Santoso, “paying for child pornography images may aggravate an offence because it can reflect on the strength of an offender’s motivation to possess the material”. We noted that offering payment for the child abuse material could also feed into the wider harm given that the payment supported the market in producing child abuse material.

(b) Offender’s participation in a network:

(i) The respondent was a member of a network (a Telegram chat group) where pornography (including child abuse material) was shared.

(c) Duration and persistence of the offending behaviour:

(i) The offence duration was fairly long, spanning from 7 May 2020 to 22 October 2021. There was also evidence of an escalation in the level of the respondent’s engagement with the child abuse material. The respondent graduated from merely accessing child abuse material in May 2020, to downloading child abuse material into his personal devices. The child abuse material also became longer, with the longest video lasting 37 minutes and 46 seconds long downloaded on 20 October 2021, which was two days before the police raided the respondent’s home on 22 October 2021. This also rendered the Defence’s submission that the respondent was a “one-off offender who sourced for [child abuse material] out of mere curiosity and sheer boredom during the COVID-19 pandemic” [emphasis in original] unpersuasive.

64 The present offence was pegged at the high end of “low culpability” chiefly because of the sophisticated means employed by the respondent through his use of the TOR Browser to access the dark web for child abuse material and his payment for child abuse material with Bitcoin in an attempt to get full access to a dark web site with child abuse material. The present offence did not cross the threshold into “medium culpability” because there was no evidence of the respondent possessing the child abuse material with an intent to further distribute it, or with an intent to make some form of gain from it (other than his own sexual pleasure). Furthermore, his participation in a network (the Telegram chat group) was also limited in that the evidence suggested that he was a passive consumer rather than an active supporter or contributor in the chat group.

65 Thus, with the respondent’s offence calibrated at the lower end of “moderate harm” and at the high end of “low culpability”, the applicable sentencing range was three to 24 months’ imprisonment with the option of a fine in addition. We determined the appropriate starting point to be 12 months’ imprisonment. What stood out was the nature of the videos which were long and explicit, and which involved penetration as well as other explicit sexual activities with very young victims.

66 We then considered the offender-specific factors in this case.

67 We gave due weight to the respondent’s TIC charges. We noted that the similar TIC s 377BK(1) Penal Code charge (DAC-903560-2022) was already taken into account when the offence-specific factors were considered (in particular to demonstrate the period of offending of more than a year, and the sophistication of the offending in the employment of the TOR Browser and Bitcoin). We thus did not give an excessive amount of weight to this TIC charge (see Public Prosecutor v BMR [2019] 3 SLR 270 at [40]). As for the other TIC charge (MAC-901458-2022), this charge was for possession of 119 obscene films, and this offence was a less serious offence which carried a punishment of a fine not exceeding $20,000 or imprisonment for a term not exceeding six months or both. Some aggravating weight was accorded to this offence given that it was prima facie a similar type of offence (Public Prosecutor v UI [2008] 4 SLR(R) 500 at [38]) and this offence was not already taken into consideration at earlier stages of applying the Logachev-type framework for the s 377BK offence. We noted the respondent’s lack of antecedents but did not consider him to be a first-time offender given his TIC charges: Chen Weixiong Jerriek v Public Prosecutor [2003] 2 SLR(R) 334 at [15].

68 We accorded due mitigating weight to the respondent’s plea of guilt, although we also noted that he was apprehended in the course of a police raid and so was caught red-handed.

69 We nonetheless discounted the starting point of 12 months’ imprisonment to eight months’ imprisonment on account of the saving of time and resource that followed the plea of guilt.

70 We therefore allowed the Prosecution’s appeal. We set aside the sentence of six weeks’ imprisonment below and instead sentenced the respondent to a term of eight months’ imprisonment. This sentence was necessary in order to achieve the ends of specific and general deterrence, which is a key consideration in the effort to stifle the demand for child abuse material.

71 We thank counsels for their submissions. In particular, we record our appreciation to Mr Santoso for the thorough and comprehensive submissions he made before us. We derived significant assistance from the submissions.