Maintaining confidentiality, privacy, and data security is paramount for TBI research, but additional care is required when working with YSGM people because the consequences of data breach can be dire. Adolescence and young adulthood are periods when many YSGM people develop their identities and decide if and how to disclose their YSGM status (14). Identity disclosure during research has important interpersonal implications (15), and thus, researchers and clinicians must consider whether third parties might cause harm if access to TBI data is acquired (16). For example, some families may reject their YSGM teenager upon disclosure of their SGM status, possibly resulting in homelessness (17). Identity disclosure can also increase the risk of bullying and victimization in high-stigma areas (18). Unintentional disclosure might occur during TBI development and validation [e.g., having one YSGM person reach out to a closeted YSGM person via snowball sampling; (19)], content design (e.g., using rainbow images in an app's logo design), content storage [e.g., content stored on the cloud vs. locally on a YSGM person's smartphone; (20, 21)], or implementation (e.g., providing the TBI at a clinic where the YSGM person is out to the provider but not the parent).

In addition to best practices in maximizing confidentiality, privacy, and data security [e.g., anonymity, encryption, data held separately from identifiers; see (22)], we recommend the following. 1.

Decide on the age range of the intended end user (23), as some issues related to unintentional identity disclosure can be avoided with older YSGM people.

Non-affirming parenting practices (26), peer bullying (27), anti-YSGM language in public spaces (28), and anti-YSGM laws (29) disempower YSGM people. Because disempowering experiences drive YSGM health disparities (30), it is important to center empowerment and autonomy in TBIs. One method is to design TBIs that adequately function if YSGM users opt-out of multiple features and data collection requests. TBI developers could also provide opportunities to promote autonomous decision-making by designing modular TBIs [vs. linear and constrained; (31)], using visually based and easy to comprehend informed consent materials, and eliciting informed consent throughout the TBI (23). Another method might be featuring YSGM people as co-authors or co-developers of the TBI. For instance, if focus groups or advisory boards are used throughout TBI development and implementation as recommended in this perspective, YSGM who participate in such activities could be featured somewhere on the TBI explaining their contributions. This representation could convey a message of “for us by us” which, in turn, could help other YSGM people feel seen and valued.

In the United States, most TBIs are classified as low-to-moderate risk by the Food and Drug Administration (i.e., the nation's regulatory body for healthcare products), so they receive minimal premarket clinical testing (32). In this regulatory environment, venture capitalists are investing heavily in TBIs [$637 million in 2019; (33)]. As a result, TBIs may enter the market without adequate effectiveness evidence, leaving the burden of determining effectiveness on individual providers (16). Furthermore, the regulation of artificial intelligence in healthcare, including TBIs, is a moving target (34). These regulatory realities raise concerns about the extent to which TBIs are based on scientific evidence.

Regulatory limitations also underscore the potential for harm in rapid TBI development and implementation. YSGM people, given their increased vulnerability in healthcare settings relative to heterosexual and cisgender people (35), require extra care to reduce the potential of “technology-facilitated abuse” (36)—that is, the harm that derives from TBIs with poor validity, accuracy, safety, and scientific rigor. Even if a local jurisdiction does not require extensive effectiveness evidence, we charge TBI developers and providers to implement TBIs with YSGM people if and only if: (a) the TBI is technologically stable, has high fidelity, and yields clinically meaningful results (37); (b) implementation stakeholders are educated on the scientific evidence behind the TBI; (c) the TBI has a robust monitoring and evaluation plan (38); and (d) the TBI presents accurate claims of its effectiveness for YSGM people.

Without cultural sensitivity and tailoring, TBI uptake and effectiveness among YSGM people may be suboptimal. Research suggests that culturally tailored interventions tend to be effective among YSGM people (39–41), especially for non-White YSGM (42) and for outcomes like HIV (43, 44). From an implementation perspective, clinical practitioners who lack understanding of YSGM issues and concerns (45) may hinder TBI uptake. Thus, we caution against adapting extant TBIs to YSGM people without adequate consideration of how intervention components are delivered in a YSGM-affirming manner. Instead: •

Involve YSGM people at all stages of TBI development and implementation to reduce the chances that a TBI perpetuates stereotypes, promotes discrimination, or causes unintended harm;

Rates of STIs differ by gender identity and sexual orientation (52), necessitating the development of differentiated interventions for YSGM people. Moreover, majority group researchers, clinical providers, or public health officials may treat YSGM communities as a monolith even though members have intersectional identities (53), unique constellations of stressful experiences (54), and diverse healthcare experiences (55). Because YSGM people are a historically neglected population (56, 57), it can be tempting to overlook the heterogeneity of this group when developing TBIs for STIs and emerging infections. We encourage the careful consideration of whether the mechanisms of the TBI are broadly applicable to YSGM people, the execution of thorough acceptability and feasibility studies with YSGM subgroups, the matching of YSGM people with subgroup-specific interventions (58), and the adaptation of existing TBIs to specific subgroups [e.g., transgender youth; (1)].

TBIs should prioritize the long-term well-being of YSGM people with sustainable plans for TBI maintenance and appropriate connections to other forms of care. Developing, releasing, and forgetting a TBI is unacceptable because YSGM people constitute a vulnerable population. For instance, with an infection like HIV, it could be useful to design TBIs to connect to wraparound services. Furthermore, YSGM people may perceive TBIs for STIs and emergent infections as spaces to seek help for comorbid concerns, especially since YSGM-specific TBIs for mental health conditions are relatively lacking (1). Because suicide is alarmingly high in this population (59), careful attention to how high-risk suicidal behaviors may be disclosed in interactive components (e.g., forums, web chat) is warranted (60, 61). We recommend creating TBIs that have clear pathways to affirming healthcare providers, easy to access crisis lines, and ongoing plans for monitoring risk if interactive components are featured (e.g., passive-sensing suicide risk detection systems that alert clinical team members to perform a suicide assessment). For emergent infections, we encourage TBI developers to design technologies that connect to other forms of care (e.g., another TBI for sexual health) in the event that an infection is rapidly controlled.

We also recommend regular technological and content updates. Technological updates are important to ensure the TBI syncs with operating systems and does not pose a safety risk. Larger technological updates may be required if the TBI technology becomes substantially outdated or no longer meets user expectations. Regular content updates are also crucial. Information on STIs and emerging infections can rapidly develop, so it is important to have content that reflects new recommendations and science.

YSGM people who might most benefit from TBIs for STIs or emergent infections (i.e., YSGM people in high-stigma, low-resource areas) are also least likely to have the TBI developed with them. Many TBIs for YSGM people are developed in urban centers where YSGM people cluster (1, 44). Even in urban areas, facilitators of and barriers to healthcare resources within the YSGM community is not uniform (62). Thus, although it may be practical to roll out a TBI in urban centers with YSGM people that are easier to recruit (e.g., gay men presenting regularly at a sexual health clinic), we recommend that TBI stakeholders have a plan to scale up distribution of the TBI to YSGM people in high-need, underserved areas. Moreover, through the inequity of the technological landscape, access to broadband, Wi-Fi, and other technological pre-requisites to receiving care are inevitably variable. In response, adaptable interventions that can remain versatile to the access barriers of a given region become essential.

While TBIs offer solutions to address the disparities in STIs and emerging infectious disease outbreaks, it is important to acknowledge and minimize the potential harm in their implementation. Many TBIs inevitably increase the screen time among users. Previous research suggests that increased screen time is associated with poorer mental health, increased behavioral problems, worsening sleep quality, and other outcomes including poor academic performance (63, 64). Therefore, TBI development and implementation must consider the potential downside of excessive technology use. Similarly, there is a growing trend of integrating gamification and serious gaming in health and wellbeing promotion in recent years (65, 66). While these approaches could generate positive behavioral changes and cognitive outcomes, excessive use of such elements could lead to addiction-like behaviors, contributing to the already existing mental health inequities among YSGM. These considerations are particularly important given the recent statement from the U.S. Surgeon General on social media use and its impact on adolescents' mental health (67). Therefore, it is crucial for TBI developers and clinicians to strike a balance between optimizing TBI engagement for positive behavior changes and maintaining a moderate amount of technology use.