A/HRC/35/9 removal of unlawful content, investigation and action against perpetrators, and the provision of redress and reparation to victims. Any such measures should comply with international human rights norms and standards; in particular, no action may amount to undue restrictions on freedom of expression. 89 38. Inputs received from some States for the report outlined measures implemented to combat online violence against women, including education, legislation, preventive actions, reporting mechanisms and various other initiatives. 90 A number of States have examined how they can use or clarify existing laws to address online violence against women, while others have adopted laws specifically addressing online violence, often in a gender-neutral manner. 91 However, reports indicate that in many States, law enforcement agencies and courts are failing to take appropriate action in situations of online violence against women, or are using such laws as a pretext to restrict freedom of expression. 92 39. Business enterprises have also begun to address this issue, for example by developing tools that allow users to block specific individuals, safeguard their privacy or tailor their interactions to protect themselves against abusive behaviour. 93 It is important that such tools comply with the requirements set out in article 19 (3) of the International Covenant on Civil and Political Rights regarding permissible restrictions on freedom of expression. Inputs to the report provided examples of community-led initiatives, including online victim support platforms and organizations offering digital security guidance and counselling.94 Other examples of ICTs designed to prevent offline violence against women include mobile device-based safety applications, such as “panic buttons”. 95 Other technology-based tools in this area include “data escrows” (third party data trusts) to facilitate confidential reporting of sexual assault. 96 D. Emerging issues: data-driven technologies 40. The advent of big data and artificial intelligence may have an impact on women’s rights and on the gender digital divide. Data-driven technologies may provide new opportunities to solve societal problems and perform a range of complex tasks in everyday life, but there is also a risk of increasing disparities for those who do not have access, and of reinforcing, or even amplifying, gender inequalities due to data gaps and bias. For example, while big data analytics may offer possibilities to make gender-based discrimination more visible and to quantify women’s political, economic, social and health 89 90 91 92 93 94 95 96 12 In accordance with article 19 (3) of the International Covenant on Civil and Political Rights, any restriction on the freedom of expression, to be legitimate, must be provided by law and necessary for the respect of the rights or reputations of others or for the protection of national security or of public order, or of public health or morals. Any restriction must be sufficiently precise to provide individuals with adequate guidance and publicly accessible, and must not confer upon authorities unfettered discretion for the restriction (Human Rights Committee, general comment No. 34, para. 25). See also GenderIT (2014), “End Violence Research Findings”; DLA-Piper, “Online harassment: a comparative policy analysis for Hollaback“ (2016); Internet Governance Forum 2015: Best Practice Forum on Online Abuse and Gender-Based Violence against Women (2015). The European Agency for Fundamental Rights has raised the issue of whether current legislation on violence against women is well-suited to fight cyberstalking and cyberharassment (see Violence against Women: An EU-wide Survey (2014)). See Association for Progressive Communications, “From impunity to justice: domestic legal remedies for cases of technology-related violence against women” (2015); World Wide Web Foundation, Web Index (2014), chap. 4.2; Association for Progressive Communications, “Technology-related violence against women” (2015), pp. 3-4. See input by Twitter #PositionOfStrength programme. In February 2017, Google released a machine learning tool designed to flag “toxic comments” (see www.perspectiveapi.com/). See, for example, inputs by Digital Rights Foundation, Access Now and Hollaback! See also www.takebackthetech.net and https://tacticaltech.org/projects/security-box. See the Centre for Internet and Society and the Bachchao Project, “Evaluating safety buttons on mobile devices: technological interventions, personal safety, and women’s agency” (2017). Amnesty International has developed a basic panic button application in an attempt to address privacy and security concerns (see www.panicbutton.io). See www.projectcallisto.org.