Cosmetic Surgery Effects Pre-Read

Raw text

Transcribed text as recorded, including headers and formatting fragments.

Exhibit 72 a? 1 I 5 1" Eh. I.) Cosmetic Surgery Effects Pre- ReadJmg pH 5 a, t I C'C V srER Decision HAHCERY c' . . . . "\M'xmfl Should we continue to prohibit the creation of Cosmetic Surgery Effects by Developers on Spark C. & H ' AR? TL;DR Cosmetic Surgery Effects change the shape of a person's face or facial features in a way that would require cosmetic surgery in real life (examples). Our ban on Cosmetic Surgery Effects was instituted in October following negative press coverage, questions from regulators, and growing concern from experts about the impact of such effects on user wellbeing and mental health. While robust, longitudinal studies of the impact of Cosmetic Surgery Effects are not available, and global wellbeing experts have expressed significant concerns to us about the impact of these effects on body dysmorphia and eating disorders. On the other hand, some users and AR creators would like to see the ban lifted, and none of our competitors have similar limitations. Although data is limited, the ban on Cosmetic Surgery Effects does not seem to have impacted key user engagement metrics, but it is possible that we will see impacts in the long run. On balance, the XFN recommends continuing the ban and erring on the side of protecting users from potential mental health impacts. The XFN proposes revisiting the issue in October on the one-year anniversary of the ban, when more data may be available. Scope A decision may impact Spark AR and its supported apps: Instagram, Messenger, and Facebook. Policy Background ln October 2019, there was a series of articles skewering lnstagram about third party Cosmetic Surgery effects. The articles focused on the negative impact these effects could have on mental health and body dysmorphia issues for teens, especially girls. Leads from the Policy, Instagram, and Spark product teams agreed to a temporary policy to stop approving Cosmetic Surgery Effects. (See background do_c_:). The temporary policy was met with very positive pass, but the temporary nature of the policy was not clearly conveyed in the coverage. Company spokespeople, including lG Head of Product Vishal Shah, have spoken publicly about the policy as part of lG's well-being efforts. (See coverage here). CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-IG-AG-00170733 In November, the Content Policy team aimed to finalize the policy for AR Cosmetic Surgery Effects through the Product Policy Forum the company's regular policymaking process. During that process, stakeholders from Content Policy, IG public policy. Spark AR product. lG product, Comms, Operations, Research, and Product Marketing generally supported continuing the ban on Cosmetic Surgery Effects. This policy development process explored various options, including age-gating Cosmetic Surgery Effects and removing Cosmetic Surgery Effects from recommendations surfaces. (For more info on the options considered and the pros and cons, see Appendix I). Ultimately, the ban was provisionally passed, pending completion of further APAC-focused data analysis, additional Outreach to experts, and additional research. Following this additional work, we decided to escalate the issue to App Leads to ensure alignment with the recommendation to continue the ban. (Apart from Cosmetic Surgery Effects, we also prohibit filters that promote racist stereotypes, including skin lightening and effects that leverage racist stereotypes, as well as graphic nip/tuck lines that imitate markings from plastic surgeons. On the other hand, we permit effects that imitate make-up or any effect that transforms the face in service of a character or a fantastical transformation such as an animal or a halloween costume. We also permit users to upload images via other photo manipuIation services that mimic cosmetic surgery. For more information, see Appendix If Examples of Types of Effects) - User Engagement & Cosmetic Surgery Effects Ban Although data are limited, the ban on Cosmetic Surgery Effects does not seem to have topline impacted user engagement metrics. It is possible that we will see impacts in the long run. This chart shows DAP for Spark AR Effects on IG before (green) and after the ban (purple): hilly/V 4t; MA {INN/W .' J A I'm'dlkr Set) 'l 9 Ci: "t 9 Nth! '19 MF- FM) '23 (For additional data, see Appendix til Additional Data.) - CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-OO170734 Potential Mental Health Impacts Independent experts we conSulted from around the world. including in the APAC region. generally agree that Cosmetic Surgery Effects raise significant concerns related to mental health and weltbeing. especially for teenage girls. (For info on what experts told us. see Appendix IV External Engagement Review.) - That said, studies that provide robust, long term data on the effects of these filters do not yet exist and they likely will not be available before the potentially damaging impact to user wellbeing manifests. The emerging evidence is concerning. as detailed in Appendix V - Research Review. Our literature review revealed that the increasing use of face filters on social media leads to unrealistic beauty standards. mental health issues, and has been postulated to be related to a rise in plastic surgery, a phenomenon often referred to as 'Snapchat Dysmorphia' (CNN health, 2018; Fardouly, 2015; The Royal Society for Public Health, 2018; The Guardian, 2018). Evidence suggests that the connection between use of photo filters and unrealistic beauty expectations is a rising global issue, including China, South Korea, the UK, and the US. An article on the impact of filters on body image (Twigg. 2019) cites an alarming rise in plastic surgery among Chinese people under the age of 28 that may be related to beauty expectations dn'ven by social media use and specifically the use of filters. Comms and Policy Risks We do not foresee significant comms or policy risk if we reaffirm the ban on Cosmetic Surgery Effects. We do foresee significant comms fallout if we reverse course and lift the ban. Also of note, some creators contend that this ban limits their artistic expression. However, after hearing the rationale behind the interim policy (user well being), many creators were accepting of the policy. For example, after a one-on-one session with top creator floly Mania to explain the policy background and rationale, that creator then went on to create other popular (non-cosmetic surgery) effects. «EN D CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-IG-AG-00170735 Appendix | - Options for'Consideration 1: Extend the Ban & Revisit in October < RECOMMENDED Finalize policy to reject AR Cosmetic Surgery Effects. o Draft policv IanQLQgg: "Reject effects that distort the user's face in a way that can only be achieved through cosmetic surgery and is not in service of creating a character transformation." Revisit in October (1 year after the temporary ban): o Review of new literature and studies on the topic of AR filters and wellbeing o Analysis of Spark growth metrics, particularly in the APAC region o Additional interviews with Spark creators, particularly in the APAC region Benefits: o Addresses concerns from safety and mental wellbeing experts Ban has been well received externally since October launch by media and experts Positions Facebook, Inc. as industry leader for long term well being of users Provides creators clear guidelines for design Commitment to revisit the issue addresses concerns that we don't have strong data to inform this decision Risks: o May be cut of line with other AR mask offerings and some users' wishes o Potentially stifles voice of creators and users o Limits possible effect types we could make as first party 2: Age-gate Cosmetic Surgery Effects Only allow users 18+ to use these effects on Facebook and lnstagram platforms Benefits: o Prevents minors one of the most at-n'sk groups - from using the filters o Allows creators to be artistic in filter creation Risks: o Does not prevent other atrisk users from using the filters (women, people with histories of body image issues) o Age-gating is not perfect. especially on instagram where we have a small percentage of age data; minors will still have access to the filters, especially on IG CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-00170736 o Could suggest that there is evidence that these effects impact minors more than adults. which is not what our research concluded o Negative s press risk: loosening the policy will support a narrative that the company is not prioritizing user well-being. o Most reporters/influencers who are against these effects are over 18, so wilI be able to see and publicly criticise them 3: Permit, but do not recommend, Cosmetic Surgery Effects Do not feature Cosmetic Surgery Effects in the Effects Gallery. Effects Tray, or other future recommendation surfaces. Benefits: o Reduces some virality of these effects o lnstagram is not seen as "promoting" these effects o Allows creators to be artistic in filter creation Risks: o Would not stymy virality as highest rate of virality comes from users (especially influencers) posting media with these effects, not from the Effects Gallery. o Negative PR risk: Loosening the policy will support a narrative that the company is not prioritizing user well-being Does not prevent any at-risk users from using the effects. 4: Lift Ban on Cosmetic Surgery Effects & Revisit in October Lift the interim ban and continue to review research and metrics. Benefits: o Allows maximum amount of voice and creativity for creators and users. o No operations lift to identify effects Risks: o Risking user-wellbeing as we wait for more research to be conducted and published o Negative PR risk: Loosening the policy will support a narrative that the company is not prioritizing user well-being o A sudden lift would flood the platform with offending-effects previously held in limbo under interim approach, raising visibility at the exact wrong time CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-00170737 Appendix II - Examples of Types of Effects 1. Cosmetic Surgery Effects (banned under policy) | 2. Cosmetic Surgery Effects + Decorations (banned under policy) | 3. Character transformation (allowed under policy) | 4. Exaggerated or humorous transformation (allowed under policy) Changes size or shape of facial features only | Changes size or shape of facial features PLUS adds one or more decorations | Changes size or shape of facial features PLUS adds multiple features that are inherent to a character | Drastically changes size or shape of facial features in a way that is obviously exaggerated or humorous and not for beauty purposes | Ex: flower crown, freckles, glasses, bow, hat, sparkles | Ex: cat ears, whiskers | Ex: witch nose, exaggerated eyebrows, bug eyes, face-aging (Enlarges the eyes) CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-IG-AG-00170738 ..... w a" 9: \L '9 . u flfiw 4 '3'». 4" "(9'; .m \'1 35%: M KP f"? r») ~ 3'2 490;. , . '- X." .- .V' t. 35 §~r 3': . a? %= , v?» $~ n . \ 3' J: \~. "1' (V my; no n I» . . "d X o I. TN '1 n" | v». n, , n 1 $1 '17. so BE? 'c u- a. v. . .H ,u * ' s n. {'39. n tw' «o HJ t v ~~ at ya" me 3P3 r. (,fi'lafifieé'ine Iiés)_"' g" r1! \a 3::- 35 ,u' w "r3 K4 am}; ~49!" I\,~ ma : ~ '! .v-w§: a. $3 .3." .,. 91 Q: .4. '9' at: 9'9' i g\ 5 ...-». H! I! .u TA '9 'fi &&_\ ,3 4 4')1 4%; . . 'ot 'w '; 3; $333 {i w) at" u.n i i .. -x' .{t 'n' u. H u .-~.- $3 u. \'v' as? H ... g . rt '2'. Eifx' . CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - IVIT-IG-AG-00170739 Appendix III Additional Data - AR Usage Overview (MAP): 0 lnstagram: 328.99 million 0 Facebook: 13.92 million 0 Messenger: 139.78 million MAP mm '3 book Note: This shows the usage of all AR effects (both first party and third party) across the Facebook, Inc. platform. All efiects in Messenger are first party (Messenger-created) effects and do not include third party effects. Dashboard CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-OU17074O Creator Growth pre/post policy: Query Active Creator 80 K 60 K 40 K 20 K 0 Apr '19 May '19 Jun '19 Jul '19 Aug '19 Sep '19 Oct '19 Nov '19 Dec '19 Jan '20 Feb '20 post_interim_beauty_queue creator_count pre_beauty_queue creator_count interim_beauty_queue creator_count APAC Creator Growth pre/post policy: Query APAC Creators 50 K 40 K 30 K 20 K 10 K 0 Apr '19 May '19 Jun '19 Jul '19 Aug '19 Sep '19 Oct '19 Nov '19 Dec '19 Jan '20 Feb '20 post_interim_beauty_queue active_creators interim_beauty_queue active_creators pre_beauty_queue active_creators CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-IG-AG-00170741 Consumer DAP .. Is M , ,5 I'v'. 3.." I," 3.5: I : 1;" :0 M .K I r ,2' 'v [\1' x , -../ \ I I; G [Mr] 9. 5oz: £3_ 5-1:: ?. I'M: 'L '2]. O:l 4. IS. 94m- ' - 2. not . I6. 33:: all. Dc: I} Jln 2F jttn . ~l; r; ,g- t) - APAC inlerim_bcuuly_quuuu lhn'd. DJYIVJIGI') .'.'_- .'c--,: ;-."-_-z =-' 5;: :1-3: .19. 0': in :1 :".'.'l,' 6.1-:- -- "1.1:". w: !'-."'r:-: t "I'u' '1 : :r'-.' .'.':' - 3 :u' i::;-u':._-,(- .- as. 1 2: -. -- A9AC post_mtnrim_hmul\r_qupuclhml_parlv__ddp Top Effects: Until the interim policy launched in October. 0% of Cosmetic Surgery Effects were ranked in the Top 100 effects by try-ons. Additional Effect Creation: The data shows that creators who created Cosmetic Surgery Effects and who did not stop creating did so on average more often than their counterparts who did not create Cosmetic Surgery Effects. 0 42% of creators using "face warp" (any effect that distorts the face that includes, but is not limited to, Cosmetic Surgery Effects) ended up publishing other effects, which is in line with overall creator retention behaviour, so there is no reason to believe that creators might churn at a higher rate because of Cosmetic Surgery policy limitations. 0 Cosmetic Surgery Effects Creators vs. Non 0 2nd effect: 28% vs. 18% 3rd effect: 20% vs. 8% OOOO 4th effect: 16% vs. 4% 5th-10th effect: 10% vs 7% 10+ effects: 5% vs 4% Effects Usage: On average Cosmetic Surgery Effects have a slightly higher share+save rate than third party non-Cosmetic Surgery Effects (15% vs 13%). CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-00170742 Appendix IV External Engagement Review - Full stakeholder encaqemen: summafl o Generally, stakeholders support making a distinction between effects achieved via makeup versus those that can only be achieved via cosmetic Surgery, though no line will be perfect. o These extreme cosmetic surgery effects can have severe impacts on both the individuals using the effects and those viewing the images. o Children are particularly vulnerable. however many others are vulnerable as well: those with a history of mental health challenges, eating disorders, etc. o An additional round of outreach was done with 9 experts from APAC and MENA and these stakeholders reiterated the dangers these filters have in advancing unrealistic beauty standards and impacting mental health and body image and their feedback was consistent with prior stakeholder feedback. W": 53253 2} .29 as;Lax-iiA "22.... u 5.2.". r, y 5,3 CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-00170743 Appendix V Research Review - Topline: While the research in this field is new and there is not yet robust causal evidence, experts and academic research from around the world APAC included generally agree that -- these effects are cause for concern for mental healthtand wellbeing and there's strong support for us to get out in front of the issue and be industry leaders. General literature review: o The increasing use of face filters on social media leads to unrealistic beauty standards. mental health issues, and has been postulated to be related to a rise in plastic surgery, a phenomenon often referred to as 'Snapchat Dysmorphia' (CNN health, 2018: Fardouly, 2015; The Royal Society for Public Health, 2018; The Guardian. 2018). o Facebook and Instagram use is associated with body image issues and anxiety among users and particularly amgng women and teenage girls. Airbrushed and unrealistic images of women online can lead to eating disorders and body dysmorphia (a condition linked to eating disorders and depression). This is more prevalent among women and even more sg greyalent among teenage and agglescent girls (Fardouly. 2015; Tiggemann and Slater, 2013). Men are, however, also found to be affected by photo-based activity on social media (Lonergan et a], 2018). o There is no research-based agreement on the effectiveness of warning signs on highly filtered photos. o While some experts and users recommend and support warning signs (Wang et a|., 2019; The Royal Society for Public Health, 2018), there is a growing body of evidence that warning labels (or disclaimer labels) on digitally altered photos do not necessarily reduce the risk of exacerbating body dissatisfaction and could rather emphasize "an ideal version of beauty" (Tiggemann, Brown, and Andeigerg...29_1_9.)- APAC focused literature review: The research summag provides more precise and meaningful context for understanding the state of research in this space. With the caveat that some of this research is better viewed in the context of other work, and is in many cases still preliminary, existing evidence suggests that the association between beauty expectations and social media use, the use of photo filter in particular, is observed cross-culturally. Some research suggests that the focus on physical appearance and the normalization of plastic surgery in some Asian countries may place these populations at higher risk for emotional impact from unrealistic beauty standards which can be formed in response to social media filters. o Body dissatisfaction and eating disorders are now global health concerns and studies indicate they are on the rise in Asia (Pike & Dunne, 2015). Studies indicate that young women in Korea and China report higher levels of body dissatisfaction and disordered CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-lG-AG-00170744 eating behaviors than young women in the United States (Pike & Dunne, 2015; Jung & Forbes, 2006). Data shows that as body comparison increases body satisfaction decreases. and that this occurs crossculturally.(Guadagno. 20181 o There is currently limited research-on the direct impact offiltered images on nonwestern countries, and non-female populations. However, the acute focus on beauty and disproportionate idealization of unrealistic beauty standards in some countries, such as in Korea and China. may lead to more exposure to and a larger negative impact of _ filtered image on users' mental health. An article on the impact of filters on body image (Twigg 2019) cites an alarming rise in plastic surgery among Chinese people under the age of 28. This may be related to beauty expectations driven by social media use and specifically the use of filters. According to this article, this relationship (between social media use and plastic surgery) is further exacerbated by the normalization of and ease of access to plastic surgery, as well as the glorification. and resulting de-stigmatization. of plastic surgery in Asian countries like Korea. o Additionally, the use of euro-centric filters may have a disproportionate negative impact on people of non-european descent. Companies, like Snapchat, along with other apps from Asia. have been accused of offering filters that "white-wash" images that "transform selfies into lighter, slimmer, and wider-eyed versions of users" (Donato 2016). CONFIDENTIAL PROTECTED BY CONFIDENTIALITY AGREEMENT - MT-IG-AG-00170745