According to research, sexual assault victims and their support providers want an alternative reporting system, such as mobile phone applications and online reporting platforms, in addition to formal criminal justice reporting methods (Heydon et al., 2023). There are many reasons for this. First, many applications and websites allow victims to report incidents in real-time and store their report (e.g., iWitnessed), which is crucial in situations where victims cannot report to the authorities. There may be no police station nearby for the victim to access support, such as in rural areas and in low- and middle-income countries (see Smith et al., 2019), or victims may not be able to escape their attacker and make a report, such as in domestic violence cases. Compared to requiring formal police interviews, giving victims the flexibility to report sexual assault through informal or anonymous channels can reduce barriers and increase overall reporting rates (Heydon et al., 2023). GBVxTech also stores reports for potential future use in criminal proceedings if victims choose to formally report later (Paterson, 2018).

Second, GBV victims frequently delay reporting incidents to law enforcement (Read and Connolly, 2007; Loney-Howes et al., 2022), which results in missed opportunities to promptly gather forensic evidence and victim accounts (i.e., memory evidence). Memory strength for the crime decreases with time (Ebbesen and Rienick, 1998; Gabbert et al., 2009, Stevens et al., 2022). With GBVxTech, victims can document a GBV incident soon after it occurs. Research has found that an early initial free recall attempt can maintain the accuracy and completeness of an individual’s account over time if victims and witnesses are interviewed following recommended practise (e.g., Penrod et al., 1982; Wixted and Ebbesen, 1997; Ebbesen and Rienick, 1998; Gabbert et al., 2009), including in sexual offence cases (Stevens et al., in prep.). Further still, immediate self-documentation can preserve memory accuracy over time (Gabbert et al., 2022; Stevens et al., 2022), and reduce susceptibility to misleading post-event information (Gabbert et al., 2012). This is particularly important in GBV cases because the victim’s account is often the primary, if not only, source of evidence (Kebbell et al., 2007). Actual or perceived gaps or inconsistencies in victim accounts can diminish prosecution odds, as officials may think gaps and inconsistencies signal that the victim lacks credibility (Freyd, 2004). Thus, timely documentation not only maintains accuracy (Ebbesen and Rienick, 1998; Gabbert et al., 2009) but also can serve to indirectly maintain the victim’s credibility (see Westera et al., 2011, 2013).

Third, some victims may prefer to use GBVxTech over contacting the police because the technology allows them to anonymously report the incidents to help law enforcement prevent future crimes. One example of an anonymous reporting app is SafeTTC (n.d.), where individuals can report incidents occurring in public spaces (e.g., public transport) to provide key information about local hotspots. JDoe (n.d.) and Callisto (n.d.) are also apps that allow anonymous reporting, and both use algorithms to monitor when multiple reports refer to the same perpetrator for purposes of identifying serial perpetrators.

Finally, GBVxTech may be used by those who wish to seek support services (Heydon et al., 2023) or who intend to seek legal redress through civil action rather than criminal prosecution. For example, some apps (e.g., JDoe and Callisto) provide victims with the opportunity to be contacted by a lawyer or legal advisor to discuss their options.

The completeness and accuracy of victim accounts are largely dependent on adherence to minimum best practice principles for face-to-face interviews (e.g., Read et al., 2009; Brubacher et al., 2014). These principles were recently extended to self-administered written interviews that use open questions and free recall formats (Gabbert et al., 2009, 2022). These principles set a minimum standard for interviews, and include establishing rapport and trust, providing narrative practise, setting ground rules, using open questions, and allowing for the victim’s account to be appropriately documented (i.e., in the victim’s own words, or via their ‘independent voice’; Powell et al., 2005). It is also crucial that interviews adhere to principles that protect the victim’s human rights (Murad Code, 2022). Since the emergence of GBVxTech is recent, there has been little consideration of how these technologies might adopt these evidence-based principles. Below, we discuss the core principles of interviewing, and how they might be applied to GBVxTech.

GBVxTech can be a vital tool for GBV protection, prevention, and response (Eisenhut et al., 2020). For example, a report from one mobile app, SafeTTC, aided in the arrest of an individual who was already wanted as a suspect for a separate assault (Spurr, 2017). Furthermore, globally, 83% of individuals own a smartphone (Statista, 2022), making smartphone-based technologies that record and report cases of GBV accessible and available to many members of the public, especially in the Global North. As such, GBVxTech is potentially a powerful tool to address the underreporting of GBV and strengthen the evidence base necessary for successful prosecutions. Whilst GBVxTech holds promise, its use requires diligence as improper use risks harm. Our study critically evaluates GBV reporting apps intended to aid victims in alerting authorities. Though well-intentioned, these tools require appraisal to realise potential benefits.

In the only research of its kind to date, Eisenhut et al. (2020) systematically reviewed mobile health intervention apps that address violence against women. The authors classified apps into five categories: emergency (e.g., send emergency alerts to selected contacts), avoidance (e.g., avoid potential incidents), education (e.g., increase knowledge), reporting and evidence building (e.g., report an incident during or after it occurs), and supporting (e.g., used to provide support for victims). The authors found that just under half of the apps (47%) had the primary function of offering immediate help in emergencies, such as alerting emergency contacts or the nearest police station before, during, or soon after an incident. Furthermore, only 14% of apps were categorised as reporting and evidence building apps, although the prevalence of these apps was found to be increasing over time. The review also revealed that there is a growing number of education and supporting apps, which suggests that there may be a shift towards building apps that support individuals after the incident rather than during an emergency.

We extend the research conducted by Eisenhut et al. (2020) by analysing the extent to which applications in the ‘reporting and evidence building’ category both employ best practice principles for eliciting accurate and reliable information from victims and take steps to protect the victims who use them from further harm. We focus on this category because this type of GBVxTech was developed for the express purpose of recording information about the crime to share with law enforcement should the victim decide to report the crime(s) to the authorities. Altogether, given the ethical, legal, and social ramifications of GBVxTech, there is a need to better understand (1) how these applications elicit GBV accounts from users, and (2) the steps being taken by GBVxTech developers to protect users and their data.

The core best practice principles discussed above are vital in gathering accurate and detailed accounts in face-to-face and written interviews. However, GBVxTech only recently emerged and therefore little is known about whether best practice principles are being adapted and implemented in this technology. We address this gap by critically analysing GBVxTech reporting apps that are currently in use and offering recommendations for how face-to-face interview principles may be adapted within a virtual context.

Our research has two objectives. Firstly, to evaluate the extent to which the best practice principles have been implemented in GBVxTech, and secondly, to draw on our findings to stimulate research and policy development so that the potential benefits of GBVxTech can be better realised.

The evaluation checklist for face-to-face interviews used by Benson and Powell (2015) was adapted to make it relevant to GBVxTech. Alongside the original criteria, another evaluation criterion exclusively related to technology (‘GBVxTech characteristics’) was added based on recommendations from Martínez-Pérez et al. (2015). The adaptations were discussed and finalised by the research team to produce the modified checklist that was used for data extraction (see Table 1 for the modified evaluation criteria checklist). Data extraction was completed by conducting a mock report within each application.

The modified GBVxTech evaluation checklist adapted from Benson and Powell (2015) consisted of four themes: questioning strategies, adherence to the interview protocol, investigative questions, and GBVxTech characteristics (Martínez-Pérez et al., 2015). Each of the criteria on the evaluation checklist was either coded as ‘present’ (if there was at least one instance of the criterion) or ‘absent’ (if no element of the criterion was seen) within each evaluated GBVxTech platform.

Two researchers independently assessed all potential GBVxTech search results against the inclusion and exclusion criteria for final selection within the review. Any discrepancies were discussed with the senior author for final selection.

The same two researchers extracted information from the GBVxTech using the above evaluation criteria. Additionally, two blind coders evaluated around 60% of the GBVxTech for purposes of inter-rater reliability (n = 8). Overall, there were moderate to perfect levels of agreement (< 60% Kappa) on all evaluation criteria and discrepancies were discussed to yield the results presented below.

Regarding question types, 92% of the GBVxTech platforms reviewed used open questions at least once, 69% used specific questions at least once, and 0% asked leading questions. Furthermore, specific questions were the most used question type, and only 23% of technologies opened with an open-ended invitation such as, ‘[Tell me] What happened?’ instead of a specific question. We also found that whilst most of the GBVxTech used open questions at least once, only 38% of GBVxTech used solely free recall responding methods, e.g., using a textbox to enter the crime report. Additionally 62% of all GBVxTech incorporated closed responding methods, such as drop-down menus and multiple-choice questions.

Regarding trust, only 23% of GBVxTech asked users for informed consent, 52% provided an introduction to the app/website, 85% discussed the purpose and aims of the technology, and 0% provided an option for narrative practice. Additionally, whilst only 15% of the GBVxTech reported specific ground rules instructions (e.g., to state ‘I do not know’), 92% allowed victims to view the final report before submitting to clarify any misunderstandings within the report.

We also investigated whether the GBVxTech encouraged an independent voice and found that only 38% offered more than one language option, and that 62% allowed users to report using a voice note or video/photo feature. Furthermore, only one of the applications (Kharita: Harass Map) allowed victims to view other victims’ reports. We also found that only 46% of GBVxTech offered a debrief (e.g., signposting to psychological support or advice).

Since GBVxTech data could be used in legal proceedings, we assessed which details were collected by the platforms. We found that 38% asked for the identity of the offender, 77% asked for the approximate time of the offence and the location of the incident (either pinpoint on a map or write the location), and 69% asked for the specific offence type such as sexual assault or rape. Additionally, 23% of GBVxTech platforms asked about the presence of possible witnesses, and 85% enquired about potential physical evidence (e.g., forensic evidence, weapons, photos/documents/videos of injuries) to provide corroborating evidence.

All GBVxTech briefly discussed how the data will be used within their terms and conditions, but only 77% clearly explained where data will be stored and for what purposes it can be accessed (e.g., whether the data can be forwarded to police, whether data are solely saved on the GBVxTech app or institutional server, etc.).

We also wanted to investigate the security options available within the apps and websites and found that 54% of our reviewed GBVxTech utilised a security feature, such as a password, a quick escape button from the application if someone approaches, or a disguise function that makes the app appear to be another type of app (e.g., a weather app).

The GBVxTech included in our review adhered to several minimum standards of best practice for interviewing victims about GBV incidents. For example, most of the applications used open questions, which helps to ensure that the victim’s account is captured in their own words and not influenced by suggestive interview questions (Elmir et al., 2011). Furthermore, none of the applications included in our review used leading questions. Leading questions can decrease a victim’s credibility as they can elicit self-contradictions within the victim’s testimony (Andrews and Lamb, 2016). By avoiding leading questions, GBVxTech can gather accounts that make for stronger legal evidence.

We also found that 12 out of 13 of the applications allowed victims to view their final report before submitting. On one hand, this could be viewed as a positive aspect of GBVxTech, since it allows victims to amend any mistakes or misinterpretations before submitting. On the other hand, by editing the final report before submitting, it may be argued that the evidence has been contaminated since the report is technically no longer a first account, and this could have repercussions in later legal proceedings. Only one application (Kharita: Harass Map), allowed victims to view other victims’ reports. Whilst no GBVxTech should have this feature, it is reassuring that so few applications enable reading others’ reports. Allowing victims access to others’ accounts is unethical, poses psychological and safety risks, and can potentially compromise the quality and utility of the victim’s own report. Seeing other accounts could invalidate the victim’s own experiences and dissuade them from reporting if they believe other incidents seem ‘worse’ than theirs. Victims’ memories also risk contamination if they incorporate details from others’ reports into their own (see Gabbert et al., 2003). Moreover, making reports public, even anonymised, endangers victims by enabling identification through case details. The lack of this feature across the applications we reviewed is positive, but technology creators must remain vigilant against its inclusion given the potential harms.

To address language and writing proficiency differences across users, we hoped that GBVxTech would incorporate a voice note or video feature to allow victims to report in their own words. We found that around 60% included this feature. This is crucial as it allows victims to disclose using their preferred method (text, voice note, video). In addition to addressing language barriers, voice note and recording facilities may increase user satisfaction and thereby people’s willingness to engage with the application and disclose information. User feedback on the Self-Administered Interview indicated many people prefer typing or recording answers over using paper and pen, because it is easier, more practical, and offers users greater flexibility (Gabbert et al., 2009).

The GBVxTech platforms we reviewed also effectively applied the principle of investigative questioning. Most of the applications asked investigative questions such as the identity of the offender, the time/location of the offence, offence type, potential physical evidence, etc. Since reports collected via GBVxTech may be used in later legal proceedings, it is crucial that these applications do not miss opportunities to collect investigation-relevant information. Whilst the inclusion of investigative questions in GBVxTech is positive, it is important to note that most applications used forced-choice closed response formats for these questions. Closed-ended questions can limit the level of detail obtained (Oxburgh et al., 2010) and constrain the user’s independent voice. Employing more open-ended response formats could elicit richer details from users in their own words.

All the GBVxTech platforms we reviewed explain in their terms how user data will be used. Furthermore, 77% clearly explain where user data are stored, such as whether the data are stored locally on the application. Ideally, all GBVxTech should explicitly detail data storage and access. This enables informed consent, as victims are able to actively understand how their data are being used (Heydon et al., 2023) regardless of the data protection rules in a given jurisdiction.

Whilst not all minimum standards of best practice in face-to-face interviewing are feasible within a digital format (e.g., verbal communication tactics to build rapport), some principles that would be simple to adapt are often missing (e.g., open questions, narrative practice, and ground rules).

Firstly, we found that only 23% of GBVxTech prompted the user for information starting with an open-ended invitation such as, ‘[Tell me] What happened?’ This low figure is concerning because according to face-to-face interview guidelines, a free recall account should always be obtained before asking specific questions (e.g., Achieving Best Evidence guidelines in the United Kingdom, Ministry of Justice, 2022). This is encouraged because it allows victims to give a complete, uninterrupted account in their own words (Powell, 2002; Brewer and Williams, 2017). Furthermore, more than half of all GBVxTech used closed responding methods, such as drop-down menus and multiple-choice questions. This response style raises numerous issues; firstly, these forced-choice methods limit the response options available, and therefore may not list the victim’s desired response option. As a result, victims may be forced to select an option that does not accurately capture their independent voice and experiences, or they may choose not to continue reporting if they feel their experience ‘does not count’ within the options provided. Secondly, since the victim’s account is limited to preselected response options, they cannot describe their experience in their own words. Forced-choice formats decrease the level of detail in reports, potentially reducing evidence quality for investigators. An open response format would also let victims rehearse details about the incident in their own words, strengthening their memory and any subsequent accounts they provide to legal officials.

We found that none of the GBVxTech employed narrative practice. Sternberg et al. (1997) found that children who practised providing a detailed account of a positive or neutral non-abusive episodic memory in the introductory phase of the interview gave more detailed narratives in the later recall phase, demonstrating the importance of practising a separate episodic recall prior to being interviewed about the event. Thus, GBVxTech can increase victim comfort and the level of detail provided by employing narrative practice. Our results also showed that most of the GBVxTech did not include ground rules instructions. This is problematic because, as mentioned previously, ground rules are useful for both adults and children during investigative interviewing to obtain both detailed and accurate accounts (Lyon, 2014; Ali et al., 2020). Future GBVxTech should consider prioritising the introductory phase of the application before a victim begins their report. During the pre-interview phase, trust in the application could be built and expectations set through ground rules and narrative practice, which will promote more detailed and accurate responses.

Another issue we found is that most applications allow users to report in only one language. The limited language options could be damaging to the victim’s independent voice. They should be able to report in their first language to capture the most accurate account (Raver et al., 2023). Moreover, if a victim is forced to report using a language in which they are not proficient, it could impact the accuracy of the report via errors in translation (Evans et al., 2019) or it could deter a victim from using the application.

Our analysis found gaps in adherence to human rights principles amongst the reviewed GBV technologies. Less than 25% requested informed consent from users, contrary to ethical investigative standards (Murad Code, 2022). Additionally, only 46% offered post-reporting support, advice, or referral to services. This lack of victim debriefing conflicts with knowledge that GBV elevates risks for mental health consequences like PTSD, depression (Tjaden and Thoennes, 2006; Campbell et al., 2009) and substance abuse (Kilpatrick et al., 1997). Best practices dictate that GBV reporting platforms should connect users with information and access to support given the empirical links between victimisation and psychological distress.

Finally, only around half of the GBVxTech we reviewed incorporated at least one security feature. This raises major ethical and safety concerns, particularly in cases where victims are reporting intimate partner violence when a partner could easily access their mobile phone or computer and retaliate (Freed et al., 2018). Therefore, security options (e.g., passwords or ‘quick escapes’) should be implemented in every app or website created for the purpose of documenting GBV incidents.

Whilst the GBVxTech we reviewed utilise some of the best practice interview principles which have been scientifically developed and tested for interviews in criminal justice contexts, we found that many of the technologies diverge from recommended minimum best practise standards for obtaining accurate and complete accounts. Digital evidence, like text messages and video recordings, have been permitted in GBV prosecutions, establishing precedent for admitting documentation from technology platforms (Hlavka and Mulla, 2018; Glasbeek, 2021). However, our findings reveal gaps in incorporating research-based memory retrieval strategies and ethical issues. Given the investigative and legal potential of these tools, further research should examine how to optimise memory recollection, ensure data protection, prioritise user wellbeing, and better align the applications to human rights-based principles. With careful design, GBVxTech could play a significant role in empowering victims whilst advancing just legal outcomes. However, additional interdisciplinary work is needed to actualise that potential responsibly and ethically.

GBVxTech should obtain informed consent, notifying users how data will be utilised, whether and how data can be withdrawn, and outlining plans if the application is discontinued (Martínez-Pérez et al., 2015; Heydon et al., 2023). Our analysis found that around half of the applications we identified were no longer accessible, raising questions about what happened to users’ data following discontinuation. Data loss risks accountability and can compromise the investigation of incidents and prevention of future crimes. As application technology evolves rapidly, many apps become defunct due to lack of funding for updates. Developers should carefully consider long-term sustainability and have transparent data protocols in case of discontinuation. Releasing applications without data retention plans risks doing more harm than good if evidence is lost when platforms cease operation.

The potential discoverability of victim disclosures on GBVxTech raises important considerations for users and designers. First, users should be advised that the information they report could be utilised as evidence in legal proceedings. Providing users with this knowledge is an essential part of the informed consent process. Second, developers must recognise that design features, like forced-choice questions, can potentially introduce memory errors and inconsistencies, which may damage victim credibility. It is critical that applications gather accounts in victims’ own words through open questions and response formats. Forced-choice options risk introducing inconsistencies that could undermine victim credibility later if the victim has to clarify answers they gave to forced choice questions. Since GBV victim testimony already faces heavy scrutiny and credibility challenges (Kelly et al., 2005), preserving free narrative is imperative.

If accounts gathered via GBVxTech are entered into legal proceedings, data authenticity and chain of custody (i.e., documentation of the sequence of handling, transfer, and storage digital evidence) will become important issues. Criminal proceedings may involve examining application metadata or obtaining sworn statements on application use. If applications allow revising responses, originals should not be overwritten, but rather preserved alongside the revised information with timestamps. Digital evidence risks manipulation, and therefore steps must be taken to better ensure data authenticity. Asking the victim the reason why they made revisions may be beneficial, as changes may otherwise imply memory unreliability at trial (Liu, 2018).

Finally, GBVxTech has immense potential to serve victims globally, but only if it is purposefully designed for accessibility across needs, resources, and contexts. The applications we reviewed would benefit from incorporating accessibility features for users with visual or cognitive impairments, such as text-to-speech and interface customisation options. Apps should partner with disabled persons organisations to incorporate accessibility best practises both in design and safeguarding features. GBVxTech should also function online, across different data bandwidths (WiFi/5G/4G/3G), across operating systems (Android, iOS, etc.), as well as offline to allow for timely documentation. Enabling offline access allows all victims to record accounts as soon as possible, benefiting evidentiary quality. Further, accessible design and clear language are crucial for GBVxTech, which likely attracts diverse users. GBVxTech should avoid legal jargon and use plain language to ensure accessibility for users with varying knowledge of GBV, the law, and reporting procedures, all of which can vary depending on jurisdiction. Our suggestion to use plain language aligns with recommended best practices for face-to-face investigative interviews (Dando and Milne, 2009; Farrugia et al., 2019). Onboard dictionaries or definitions can further aid understanding, particularly for crime classification questions, as many individuals lack awareness of distinctions between crimes like sexual assault versus sexual harassment. However, only five of the applications we reviewed offered built-in definitions. Developing a more inclusive GBVxTech platform has immense potential to broaden access to justice globally, especially if designed intentionally for clarity and transparency through built-in support features and avoidance of context-specific terminology.