The Seraph Project is a UK-based Community Interest Company (CIC) dedicated to preventing real-world physical violence using online signals and evidence. Our mission is to transform the way that online providers identify and handle threats of violence.
By advocating for data-driven identification of credible threats, the project aims to establish robust guidelines and escalation processes that mirror the high-stakes interoperability found in counter-terrorism and child safety sectors.
When a person is actively planning or poses a severe risk of impending violence, evidence of this potential risk is extremely common. Most notable is leakage, where a person communicates an intent to harm another, either to their target or to third parties. Many other signals such as evidence of preparation or a history of prior violence are also common.
As more human communication moves online, so does more of this evidence, now logged and timestamped and preserved by a range of online platforms and services, with dedicated teams of professionals devoted to safety and wellbeing. The combination of skilled specialists, huge quantities of data, and powerful tools make these platforms uniquely well placed to detect threats and pre-violence behaviors. Through this, many instances of violence are not just predictable - they are preventable.
There is strong evidence that this approach can have a substantial and meaningful impact. For a few specific areas such as Terrorism and Child Safety, strong mandates and processes are backed up by dedicated organizations and resources like The National Center For Missing and Exploited Children and the FBI Cybertipline. These powerful interoperabilities ensure that technology companies have a clear and well supported structure for taking action to prevent violence, while also creating strong incentives and encouragement for them to do so.
Unfortunately, for most violent actions this opportunity is nowhere close to being fully realised. High quality resources and support sustems either do not exist or are extremely limited in scope. Technical challenges, business pressures and a lack of public and regulatory attention combine to create significant incentives to neglect, deprioritize, or misclassify these risks that cost people their lives every day.
All of this means that the critical dangers that lead to a significant majority of global homicides are chronically undersupported, both by online platforms and by regulatory authorities and law enforcement.
We want to change that.
These areas represent our immediate priority as they account for the majority of personal-cause homicides. Because these events involve pre-existing relationships, there is also an increased likelihood of available and actionable data, making these cases the most viable candidates for predictive intervention.
Our long-term aim is to expand this coverage to include all forms of individualized kinetic violence.
The Seraph Project is led by James Gresham who has worked for 15 years in roles focused on Trust & Safety and Data Science & Analytics. In addition to his primary work on direct harm prevention, he has worked across a broad range of operational areas related to content moderation, user safety, and platform experience at companies including Alphabet, Meta and Shopify. As an active member of the Trust and Safety community James has served as an advisor for the Trust & Safety Professional Association (TSPA), as well as writing educational content for the TSPA curriculum and other educational organisations.