What is Trust & Safety?

Trust & Safety (T&S) refers to the work of teams within digital platforms that are responsible for defining, detecting, and mitigating harmful online behaviours. The Trust and Safety Professional Association (TSPA) defines it as “…an umbrella term to describe the teams at internet companies and service providers that work to ensure users are protected from harmful and unwanted experiences.”

TSPA lists the following functions commonly performed by Trust & Safety professionals:

• Content policy
• Content design and strategy
• Data science and analytics
• Engineering
• Legal
• Law enforcement response and compliance
• Operations
• Product policy
• Product management
• Public policy and communications
• Sales and advertiser support
• Threat discovery and research

It is important to understand that not all Trust & Safety teams are set up in the same way, but the areas listed above broadly represent the kind of work that Trust & Safety is about.

Why Trust & Safety Matters

Trust & Safety is important because it fills a crucial role in the threat response ecosystem. We ought to think about the response to the problem of online harms at a systemic level. On one end, we have legislative responses (laws and regulations). These are “hard” responses, meaning breaches of such laws and regulations can be enforced, potentially even with criminal penalties.

On the other end, we have “soft” responses, such as public education. These are important because some harmful things will slip through the net and reach the intended targets no matter what. But there is no guarantee that education alone will be effective.

Trust & Safety can be thought of as a middle ground between these two ends of the system. Trust & Safety performs functions that are, in some ways, similar to governments. The creation and enforcement of content and product policies are similar to the ways in which governments and regulators create laws and regulations, albeit Trust & Safety policies are enforced by Trust & Safety professionals and not by public authorities. These responses are not as “hard” as legal enforcement, but they have more bite than mere public education campaigns.

Trust & Safety as a response to online harms has two main advantages: speed of response and the ability to proactively prevent harm.

• Speed of response

Government action is powerful but often slow. It may also happen too late. Passing laws takes time. Enforcing laws also takes time, often requiring court proceedings. Broadly, governments can do two things:

• Pass laws regulating online harms, which tend to be broad and high level.
• Enforce laws, which is almost always retrospective: something needs to go wrong first, be reported, investigated, and brought to court or a regulator.

This may be too slow for a victim of, for example, doxxing or persistent online harassment.

Often, the quickest and most effective remedies lie with technology platform providers. A user who is being harassed on a social media platform does not really need a legal remedy if the harasser can simply be removed from the platform. Even before victims go to the police or regulators, their first port of call is usually the platform’s complaints mechanism, which is typically under Trust & Safety.

• Ability to proactively prevent harm

Technology platform providers also have the advantage of preventing harmful content from reaching users in the first place. Trust & Safety policies and procedures can define, detect, and remove harmful content even before it reaches its intended target.

It is therefore encouraging to see large platform providers actively developing their Trust & Safety capabilities. These measures include both the deployment of technological countermeasures as well as investment in research and training.

Take Google, for example. On the technology front, in 2024–2025, Google reported that over 92% of human reviews for harmful apps on the Play Store are now AI-assisted, significantly reducing the time a malicious app remains available.

On the research and training front, in October 2025, Google announced it would award $5.6 million in funding to 84 researchers as part of a push to advance AI safety. The grants, awarded through the Google Academic Research Awards (GARA) program, support researchers working on AI safety, addressing technical challenges, and ensuring responsible AI development.

Trust & Safety as an Emerging Profession

It is in the interest of the whole ecosystem—governments, civil society, and individuals alike—to support these efforts to prevent online harms upstream. Prevention is, as the saying goes, better than cure.

Given the importance of Trust & Safety as a function, professionals need to be equipped with the right knowledge and skills. Those involved in policymaking and enforcement perform functions similar to public service and law enforcement officers, making skills from those backgrounds particularly valuable.

Similarly, reviewing cases, making decisions, and handling appeals can draw on lessons from the legal system, while professionals engaged in research could benefit from skills such as futures thinking and horizon scanning. Everyone working in Trust & Safety would also benefit from a solid grounding in ethics, as the field fundamentally revolves around treating people right.

There are broader intellectual questions to consider as well, especially when Trust & Safety policies intersect or clash with local governments. This calls for an understanding of political and constitutional concepts such as legitimacy, authority, and democratic mandates.

Crucially, beyond skills and knowledge, Trust & Safety professionals should share a common set of values and be clear about why they do what they do and whom they serve—much like the shared principles encapsulated in the Hippocratic Oath for doctors.

As noted earlier, there is currently no universally accepted definition of online harms, leaving individual companies to set their own terms. This creates uncertainty and gaps in the system, where actions deemed actionable on one platform may not be on another, potentially harming real people simply because something fell through the cracks. Establishing a shared set of values could go a long way toward addressing these inconsistencies.

Challenges and the Road Ahead

At present, Trust & Safety professionals are drawn from all sorts of backgrounds. Unlike law or medicine, there is no common educational pathway. There is no guarantee that professionals will be equipped with the skills, knowledge, or values they need to do their jobs effectively. This is not ideal—for professionals or for the people they are trying to protect.

The problem of online harms will only grow more acute. That means Trust & Safety professionals will be more important than ever. It takes time for a new profession to build its culture, history, and institutions.

An analogy may be drawn to the legal profession. This year marks the 200th anniversary of the Second Charter of Justice, the foundational document of modern Singapore’s legal system. The legal profession is an old one, by Singapore standards, yet ethics and values training has taken a long time to become systematised.

The requirement for formal ethics training in law schools was only implemented in 2024. This was done reactively, in response to an incident in 2020 which uncovered cheating by students taking the Bar Exam, the professional qualification examination for Singapore lawyers.

We do not have the luxury of letting Trust & Safety develop at the same pace, nor should it develop reactively. When online harms occur, they occur quickly and spread rapidly.

For example, “cancel campaigns” can quickly emerge online, targeting individuals and leading to harassment and distress. One instance was the #PunishXiaXue campaign, in which local influencer Wendy Cheng, known as “Xiaxue,” was targeted for allegedly making racist remarks. The petition against her garnered over 23,000 signatures.

As a public figure, Xiaxue was able to handle the backlash, but not everyone can.

Consider the case of Mark Lin Youcheng, a volunteer at a dog shelter who witnessed a hit-and-run incident involving one of the shelter’s dogs. Frustrated that the driver had not taken responsibility, Lin photographed the car’s licence plate and posted it online to identify the driver. The vehicle was traced to Soon Kim Choo, after which Lin posted her contact details online and urged others to “give her hell,” including on her employer’s Facebook page.

In reality, Soon had lent the car to someone else and had not been driving that day. Lin was later convicted under the Protection from Harassment Act 2014.

Legal remedies for online harms often come too late—the harm is already done. Proactive protection of the public is required, and the people best placed to provide it are Trust & Safety professionals working with platform providers.

Technology safeguards alone cannot be trusted to provide an adequate level of protection.

Consider issues like “Cancel Campaigns”: there is a fine line between exercising one’s legitimate freedom of expression and harassment and bullying. Where is the line to be drawn?

It requires context and sensitivity—in other words, it is a value judgment.

Such judgment cannot be provided by technology alone, which relies on rigid metrics to make decisions and lacks moral agency. Important and complex decisions require a human mind that is capable of subtlety, moral agency, and, perhaps most importantly, compassion.

This is why, despite advances in technology, it is more important than ever to ensure that the humans in the loop are properly trained and equipped.

The challenge is urgent, and therefore the Trust & Safety profession must be built up as quickly as possible. One way to accelerate this process is to learn from other disciplines, such as law, which have already had to confront similar challenges.

This is one reason why Singapore University of Social Sciences has partnered with Google to develop a first-of-its-kind course specifically for Trust & Safety professionals, drawing from a multitude of disciplines including law, policymaking, and criminal investigations.

Online harms are a threat to all of society, especially to vulnerable members like children and the elderly. Investing in a stronger Trust & Safety profession, as a key component of the online safety ecosystem, is in everyone’s interest, including platform providers: safer platforms serve users better. Professions arise in response to need—for example, the need for medical experts to help the sick. The need for a response to online harms is clear, and the Trust & Safety profession must rise to the challenge.

The views and recommendations expressed in this article are solely of the author and do not necessarily reflect the views and position of the Tech for Good Institute.