Emerging technologies – trends, challenges and opportunities
- Increase in use of cloud storage, encryption and smartphones.
- Encryption is the biggest challenge.
- Artificial Intelligence, investigation software, and sharing of data/intelligence is most helpful.
- One in five police officers have used AI tools in their investigations.
Increase in use of cloud storage, encryption and smartphones
The police officers reported that major technology trends that affect methods for production, storage and distribution of child sexual abuse material are increased use of cloud storage, encryption of a range of devices and online storage spaces, and the development and spread of smartphones.
An increase in use of cloud storage was mentioned by more than one third of the surveyed police officers. Increases in encryption and smartphones were mentioned by one fifth of the respondents. This includes encryption of apps, mobile phones and computers as well as encryption of cloud services.
Darknet, social media and online gaming
More than one in ten respondents answered that increased use of TOR/darknet (16.1 %), and social media and online gaming (14.3 %) being used to produce and share child sexual abuse material are major trends.
Apps, link sharing and P2P
Other trends mentioned were development of mobile applications, and an increase in offenders sharing links to online spaces (often cloud storage) or websites where child sexual abuse material can be downloaded, rather than sharing the material itself directly. Increased use of P2P networks was also mentioned.
Most frequently mentioned emerging trend in technologies that are used to produce, store and distribute child sexual abuse material*
“ Use of cloud storage appears to be increasing year on year with offenders storing and sharing links. Devices are often forensically cleaned making detecting offences more difficult.”
“ Distribution of links which lead to cloud storage accounts containing child sexual abuse material.”
“ Cloud storage is quickly going to overtake mobile phone storage (if it hasn’t already). Every case seems to have a Dropbox, Drive, or Mega connection to it.”
“ The use of cellular devices to store and distribute child sexual abuse material across a myriad of Apps and the storage of that material in third party cloud storage companies.”
“ Apps that allow covert and encrypted communication between offenders and children.”
“ Live-streaming, being recorded using DU recorder or ITube. Then shared via encrypted chat plaforms like WhatsApp and Telegram between offenders. Still large use of Dropbox links being shared on chat groups.”
“ Cellular/mobile devices and tablets/laptops are loaded with encryption by default which makes processing the device as well as their backups increasingly difficult.”
“ Devices that automatically encrypt data. Devices that auto-lock down and require passcodes to gain entry, such as smartphones. Apps that have “My Eyes Only” type secure vaults to hide content, such as what Snapchat uses.”
“ Production is made easy by the use of smartphones and distribution is, in my experience, made easy by the use of TOR.”
“ Cloud sharing. Perpetrators are locating victims through gaming apps a lot more these days.”
Breakdown of the types of encryption specifically mentioned in the survey
Encryption is the biggest challenge
Nearly half of the surveyed police officers reported that encryption is the biggest challenge they face in child sexual abuse investigations. Fourteen percent (14.4 %) of the respondents specifically answered that encrypted and locked smartphones and computers are particularly challenging.
Darknet and cloud storage
Nearly one fifth of the respondents mentioned TOR/darknet and cloud storage as technologies that are particularly challenging. In relation to cloud storage, several respondents specified that the major challenge is jurisdictional issues, for example when the cloud host is based in another country. Other challenges mentioned were live-streaming and apps that don’t store content, VPN, ISPs that don’t track IP addresses and store information, and apps with user anonymity.
“ The problems associated with storage on international servers and the legal challenges associated with that. Also, I have seen a lot of sharing via anonymous file sharing websites, especially on the ‘dark web’.”
“ The standard use of hardware encryption on smartphones and computers.”
“ Encryption/password protection Apps that store very little data on the device i.e. Snapchat.”
“ VPN services are a particular challenge as they mask the true location where someone is accessing data from.”
“ The vast amount of programs that enable a user to show live videos and keep no records has allowed offenders a larger group of children to groom and exploit.”
“ The biggest hurdle outside of encryption are ISPs which do not retain the needed data to identify subscribers. Many cellular providers do not maintain records of session and login activity. Those that do may only keep them for 60–90 days, which is often outside the range of the abuse when reported.”
“ That is the co-op with the social media platforms. They are not that willing to help the police, and that is a big problem. Use things like letters rogatory/MLAT*, and that result in months and months of waiting, and then the evidence is gone.”
“ The biggest challenge is that the use of the DarkNet is constantly rising and that using mobile applications makes it harder for us to track and search for these kind of material.”
Technology developments that police officers see as particularly challenging to child sexual abuse investigations. Share of respondents that have answered a specific challenge*
Artificial intelligence, investigation software, and sharing of data/intelligence is most helpful
Half of the surveyed police officers answered that different types of technology developments, such as Artificial Intelligence (AI), Face Recognition, PhotoDNA(1) for video and forensic tools for breaking locked smartphones are most helpful to child sexual abuse investigations. Most of those, one fifth of the respondents, answered that AI is helpful.
“ AI for image detection.”
“ Facial recognition, AI.”
Software and sharing data/intelligence
One third of the respondents named specific software for investigative work and analysis. One quarter mentioned the importance of sharing data and growing different intelligence sources like databases of hashes, and work done by NCMEC(2), Project VIC(3) or CAID(4).
“ Continued growth of MD5(5) hash sets of known material. More importantly the growth of PhotoDNA of Known Images. Most important the use of Photo DNA on movies so that non matching MD5 values on movies can be addressed.”
“ Hash databases that are contributed globally.”
“ NCMEC/CVIP(6) are very helpful.”
“ CAID/Project VIC.”
Platform and service providers
One in ten answered that it would be helpful if ISPs and social media platforms collaborated more with law enforcement. Just over three percent of the surveyed police officers said that they were helped by just that, improved help from social media companies.
“ Online storage and businesses that monitor online storage for known hashes.”
“ The implementation of any tools used by any ISP/ESP that recognizes child exploitation from the get go and is then reported to the appropriate jurisdiction.”
“ Constant cooperation between law enforcement and internet service providers.”
No helpful developments
Nearly seven percent of the respondents said that they didn’t feel that there were any technological developments that are helpful to child sexual abuse investigations.
Technology developments that police officers see as particularly helpful to child sexual abuse investigations. Share of respondents that have answered a specific technology that is helpful*
Breakdown of technologies listed as helpful in the survey
One in five police officers have used AI tools in their investigations
One fifth of the surveyed police officers answered that AI is helpful in child sexual abuse investigations. This number is also reflected in how many of the respondents reported having used AI in their investigations. The large majority, nearly eight in ten of the surveyed police officers, however answered that they have not used AI in their investigations. Several respondents commented that they would like to.
“ I would like too but haven’t taken the time to learn the technique.”
“ Would love to try AI in my cases.”
Triage and locate material
Of the respondents that used AI classifiers in their investigations, eight in ten reported that they were useful. The majority of those (46.2 %) reported that the primary gain is that the AI classifier helps to triage or filter material, to locate where to look in the caseload or highlight material that is likely to be child sexual abuse material.
“ Due to the immense amount of digital evidence, the use of AI […] allows examiners to target potential images first. This provides both a benefit in efficiency as well as a mental boost because it allows the user to sort out the most likely media files first.”
“ Everything identified by AI still has to be manually reviewed however pre-categorization of child abuse images or vetting of known system files speeds processing time greatly.
“ The image classifiers can surface potential child pornography images more quickly.”
“ It can be used as a triage tool to decide if a device needs to be deeper examined or not.”
Needs to mature
Those who reported that AI technologies are not useful, commented that the technology still needs to mature, but will be beneficial in the future, or that they had just started using it and needed to get more experience.
“ Technology doesn’t seem mature enough yet….Getting better but still a way to go.”
“ Will be more beneficial as the technology develops.”
“ I have only barely started trying to apply AI/machine learning in my cases but I’m excited to find ways to make it useful.”
Share of respondents who have used artificial intelligence technology/classifiers in their investigations
Whether respondents who have used artificial intelligence technology/classifiers in their investigations find it useful
(1) PhotoDNA is robust hashing technology used to identify child sexual abuse material. See section three of the report for a description of PhotoDNA.
(2) NCMEC (National Center for Missing and Exploited Children) is the US’ national clearing house for reports on child sexual abuse material for US based platform providers
(3) Project VIC is a global partnership between law enforcement and the private sector. It promotes data sharing between domestic and international law enforcement agencies working on cases pertaining to sexual exploitation of children.
(4) CAID (Child Abuse Image Database) is a national system to enable collaboration between law enforcement agencies working on child sexual abuse cases in the UK.
(5) MD5 is a type of binary hash. Read more about binary hashing in section three of this report.
(6) CVIP (National Child Victim Identification Program) is a database of child sexual abuse material maintained by the United States Department of Justice and NCMEC.