Introduced to tackle growing concerns over the safety of internet users – particularly children and vulnerable groups, the Online Safety Act (OSA) marks a significant shift in the regulatory landscape for businesses operating online platforms in the UK.

Passed in October 2023 and progressively being enforced, it has introduced a wide range of new obligations, imposing stricter requirements for transparency, age verification and content moderation to create a safer online environment.

Under the Act, businesses operating online must now ensure transparency by regularly publishing their safety measures and reporting on their efforts to regulators. This means not only creating new policies where needed, but also providing evidence that these policies effectively mitigate risks associated with harmful content. The Act places specific emphasis on platforms accessed by children, requiring additional safeguards and age-appropriate design features.

To comply with these new regulations, digital platforms will be required to implement more stringent risk mitigation policies and are mandated to collaborate with Ofcom, the UK’s communications regulator. Ofcom will oversee the implementation of the Act and enforce penalties for those not in compliance. To comply, businesses must maintain detailed compliance records by continuously updating and improving their safety measures to keep up with evolving risks.

Effective Age Verification and Safeguards for Children

One of the most critical elements of OSA is the focus on protecting children and young people as and when they access the internet. Come 2025, online platforms accessible to minors will be required to implement age checks to accurately determine whether or not users are children.

Ofcom will publish final guidance in early 2025, however, in the meantime it is clear that basic or outdated age-check systems – such as a simple ‘yes/no’ checkbox or self-declared age – will not suffice, and highly effective age assurance measures must be used. Innovative technologies that verify users’ ages while protecting their privacy are not a pipedream; they are available and ready to be deployed.

Platforms will also be expected to integrate further age-appropriate design features that reduce the risk of children encountering harmful content. This means filtering out explicit material, protecting personal data, and setting limitations on interactions with adults, all while maintaining a user-friendly experience. For example, social media platforms will need to assess how they moderate conversations, regulate social interactions, and structure the visibility of certain types of content.

The Need for Content Moderation and Transparency

Encouraging effective content moderation is another key element of the Online Safety Act. Businesses are obligated to implement systems to moderate harmful content – including hate speech, violence, and inappropriate material that could harm users, particularly minors. To achieve this, platforms must adopt proactive rather than reactive measures to prevent harmful content from being uploaded or spreading before it reaches their users. Content moderation efforts must also be transparent, with businesses documenting and publishing their policies, any actions taken, as well as their results.

The Act is designed to hold platforms accountable, not just for the safety measures they put in place, but also for how well the measures work in practice. Companies failing to demonstrate robust content moderation could face legal repercussions or fines from UK regulator Ofcom.

Technologies to Make the Internet Safer

Safety technology solution providers have been continuously innovating and developing solutions to keep up with the ever-changing and challenging online environment. In the age assurance space, technological advancements and the introduction of AI-driven techniques have meant that safety tech providers can now offer a range of highly accurate, privacy-preserving age assurance methods that protect user privacy, minimise friction, and ensure compliance with ever-evolving regulations.

While some methods require user interaction, such as uploading an image of an ID document or taking a short selfie video, other methods use existing user data. This data, such as an email address, can often be collected as part of the account creation process or during the checkout process on online marketplaces, and can be deployed in the background with no further user interaction required. Email address age estimation can accurately determine a user’s age without requiring sensitive personal information, allowing businesses to maintain compliance while protecting user privacy.

Within content moderation, Artificial Intelligence (AI) will play a critical role in helping platforms maintain an even safer environment. The technology can be utilised alongside human moderators to add an additional layer of support and scalability, quickly removing harmful material at scale.

An Opportunity for UK Businesses

For UK businesses, OSA is not just another regulation to follow but a hugely important opportunity to make the internet safer. By adopting cutting-edge safety measures and prioritising transparency, businesses can build trust with their users and demonstrate a commitment to protecting children when they venture online.

Businesses that proactively harness and implement effective age verification and content moderation will also benefit from the ability to avoid regulatory fines and quickly adapt to future regulatory changes. Considering the fast-paced nature of the internet, companies that are able to stay ahead of regulatory requirements now will be better positioned to thrive and grow in the years to come.

As a new piece of legislation, the OSA naturally requires businesses to change how they operate, which may initially prove challenging. However, by staying up to date on regulatory changes, leveraging cutting-edge technologies, and implementing them effectively, businesses can strategically position themselves to become a trusted voice in their space and ultimately better protect kids and young people online.

Read more:
The Impact of the Online Safety Act on UK Businesses

In today’s digital age, billions of pieces of content are uploaded to online platforms and websites each day.

Moderating this material has, therefore, never been more critical or challenging. While most of this uploaded content may be positive, we are also seeing a growing volume of harmful and illegal materials – from violence and self-harm to extremist rhetoric, sexually explicit imagery and child sex abuse material (CSAM).

Tackling this deluge of harmful content is now a defining challenge for businesses, with those unable (or unwilling) to do so opening themselves up to significant penalties and putting children at severe risk.

Our own research has revealed that over a third (38%) of parents have been approached by their kids after seeing harmful or illegal content, with many accessing materials as graphic and harmful as CSAM within just ten minutes of going online.

Therefore, the time has come for stronger content moderation measures and businesses looking beyond traditional manual moderation methods, which have become impractical and unscalable. Instead, they should leverage the complementary capabilities of AI that are transforming the landscape of content moderation through automation, enhanced accuracy, and scalability.

However, as with any new innovation, companies interested in using AI should ensure they implement the technology in a way which ensures regulatory compliance. The decisions companies make today will massively impact their future operations.

The helping hand of AI

AI has drastically transformed the content moderation landscape by using automated scanning of images, pre-recorded videos, live streams, and other types of content in an instant. It can identify issues such as underage activity in adult entertainment, nudity, sexual activity, extreme violence, self-harm, and hate symbols within user-generated content platforms, including social media.

AI is trained on large volumes of “ground truth data”, collecting and analysing insights from archives of tagged images and videos ranging from weapons to explicit content. The accuracy and efficacy of AI systems directly correlate to the quality and quantity of this data. Once trained, AI can effectively detect various forms of harmful content. This is especially important during live streaming scenarios, where content moderation needs to be viable across diverse platforms with varying legal and community standards.

While an automated approach not only accelerates the moderation process, but also provides scalability – a vital feature in an era where solely human moderation wouldn’t be possible with the sheer volume of online content.

A synergy of AI and humans

AI automation brings significant benefits, allowing organisations to moderate at scale and reduce costs by eliminating the need for a large team of moderators. However, even the most advanced technology requires human judgement to accompany it, and AI is far from being perfect on its own. Specific nuances and contextual cues can confuse systems and generate inaccurate outcomes. For instance, AI might be unable to differentiate between a kitchen knife used in a cooking video and a weapon used in an act of violence or confuse a toy gun in a children’s commercial with an actual firearm.

Therefore, when AI flags content as potentially harmful or in violation of guidelines, human moderators can step in to review and make the final call. This hybrid approach ensures that, while AI extends the scope of content moderation and streamlines the process, humans retain the ultimate authority, especially in complex cases.

Over the coming years, the sophistication of AI identification and verification techniques will continue to increase. This includes improving the accuracy of matching individuals featured in various types of content with their identity documents—a next step in ensuring consent and mitigating unauthorised content distribution.

Thanks to its learning capabilities, AI will constantly improve its accuracy and efficiency, with the potential to reduce the need for human intervention as it continues to evolve. However, the human element will continue to be necessary, especially in appeals and dispute resolutions related to content moderation decisions. Not only do current AI technologies lack the nuanced perspective and understanding, humans can also serve as a check against potential algorithmic biases or errors.

The global AI regulation landscape

As AI continues to expand and evolve, many businesses will be turning to regulatory bodies to outline their plans to govern AI applications. The European Union is at the forefront of this legislation, with its Artificial Intelligence Act coming into force in August 2024. Positioned as a pathfinder in the regulatory field, the act categorises AI systems into three types: those posing an unacceptable risk, those deemed high-risk, and a third category with minimal regulations.

As a result, an AI office has been established to oversee the implementation of the Act, consisting of five units: regulation and compliance; safety; AI innovation and policy coordination; robotics and AI for societal good; and excellence in AI. The office will also oversee the deadlines for certain businesses to comply with the new regulations, ranging between six months for prohibited AI systems to 36 months for high-risk AI systems.

Businesses in the EU are, therefore, advised to watch the legislative developments closely to gauge the impact on their operations and ensure their AI systems are compliant within the set deadlines. It’s also crucial for businesses outside of the EU to stay informed on how such regulations might affect their activities, as the legislation is expected to inform policies not just within the EU but potentially in the UK, the US and other regions. UK and US-based AI regulations will follow suit, so businesses must ensure they have the finger on the pulse and that any tools they implement now are likely to meet the compliance guidelines rolled out by these countries in the future.

A collaborative approach to a safer Internet

That being said, the successful implementation of AI in content moderation will also require a strong commitment to continuous improvement. Tools are likely to be developed ahead of any regulations going into effect. It is, therefore, important that businesses proactively audit them to avoid potential biases, ensure fairness, and protect user privacy. Organisations must also invest in ongoing training for human moderators to effectively handle the nuanced cases flagged by AI for review.

At the same time, with the psychologically taxing nature of content moderation work, solution providers must prioritise the mental health of their human moderators, offering robust psychological support, wellness resources, and strategies to limit prolonged exposure to disturbing content.

By adopting a proactive and responsible approach to AI-powered content moderation, online platforms can cultivate a digital environment that promotes creativity, connection, and constructive dialogue while protecting users from harm.

Ultimately, AI-powered content moderation solutions offer organisations a comprehensive toolkit to tackle challenges in the digital age. With real-time monitoring and filtering of massive volumes of user-generated content, this cutting-edge technology helps platforms maintain a safe and compliant online environment and allows them to scale their moderation efforts efficiently.

When turning to AI, however, organisations should keep a vigilant eye on key documents, launch timings and the implications of upcoming legislation.

If implemented effectively, AI can act as the perfect partner for humans, creating a content moderation solution that keeps kids protected when they access the internet and acts as the cornerstone for creating a safe online ecosystem.

Read more:
AI, a helping hand for businesses when moderating content

Innovation and regulation must work together to foster trust in new technologies. Innovation pushes the boundaries of what can be achieved, keeping the industry continuously moving forward.

Whereas, regulation drives the adoption of frameworks, providing mechanisms for accountability and helping organisations adapt to evolving challenges. It guides technological progress in a way that aligns with societal values and long-term goals.

On one hand, digital innovation can often happen at an accelerated pace. It can drive the development of new advanced tools and platforms online, generating new ways for people to interact and share content. However, this can create online spaces with new considerations and challenges for online safety… When environments are created without the correct regulatory measures in place, people, especially young adults and children, can be exposed to inappropriate, harmful or illegal content, or even other people with nefarious intentions.

On the other hand, regulators can be slow to catch up with innovation, often stifling ambitious ideas and hindering growth and development within the industry. Therefore, it is imperative that the two work hand in hand.

Finding the balance and practical solutions

Regulation and innovation should create an environment that protects consumers and allows businesses to thrive: one that paves the way for an online environment that’s both safe and full of exciting opportunities. It’s a complex balance to strike, but online safety should always be considered first before rolling out new tools and services that potentially put young people at risk.

While overly restrictive regulations can hinder creativity, lax rules can lead to unintended consequences, and it’s important to find the perfect middle ground. Therefore, regulators, industry leaders, and innovators must keep an open dialogue to create clear, achievable regulatory frameworks that are adaptable to new developments. Businesses should work with experts to understand and implement the available practical solutions that work best for their platform’s risk and use cases.

When developing online tools and services, online platforms and social media firms should consider the following key areas: monitoring the content they or their users produce, ensuring only those of the correct age can access age-restricted content, services, or products, and restricting or monitoring communication between specific age brackets.

Age assurance

Platforms that offer age-restricted content, services or products must implement robust age checks that clearly determine the ages of their users, ensuring they deliver age-appropriate experiences for minors while complying with legislative requirements.

Common use cases of age assurance include age-gating, whereby platforms can restrict access to content, products, or services based on a user’s age. This is commonly seen in various industries, such as alcohol, tobacco, and adult entertainment. Platforms can also focus on creating online environments with activities, content, or products suitable for specific age groups and restricting communication between certain age brackets. This is often seen on social media, video sharing, and online video gaming platforms.

There is no one-size-fits-all approach, both online and offline, and therefore, optionality is key. Businesses should work with safety tech providers to offer a range of methods that address individual preferences, minimise business disruption and mitigate the risks associated with bias and exclusion.  A range of innovative methods are available to the market to estimate age, including using an email address and facial age estimation plus verification methods, such as government-issued ID, credit card, mobile phone number, and name and address checks.

Ultimately, these technologies must be robust, low friction and privacy-preserving, ensuring minimal disruption to the user journey. Implementing age assurance measures shouldn’t cost you legitimate business, and educating users on the necessity of these measures is critical to maintaining business as usual. Clarity, a frictionless user experience and minimal business disruption will help boost trust in technology while ensuring organisations remain compliant, too.

Content moderation

In an environment where one in 10 children are exposed to age-restricted content within 10 minutes of going online, it has never been more critical for platforms to have solutions in place that can prevent, monitor and remove harmful, age-restricted or illegal content from their websites.

Content moderation solutions incorporating AI, humans or, ideally, both, can remove harmful and illegal content as soon as it is identified or, better still, prevent it from being published in the first place. This process creates better and safer experiences on platforms while ensuring regulatory compliance.

Regulators globally are actively developing guidance for protecting children from online risk, and enforcement is either live or imminent in many jurisdictions. For example, the UK’s Online Safety Act 2023 requires technology platforms to keep children safe online and ensure content is age-appropriate, legal, and within specific guidelines. Its regulator Ofcom also plans to hold websites accountable for exposure to harmful content, bringing in hefty fines if firms fail to do so.

Striking the balance

Innovation and technology adoption are happening faster than ever, and collaboration with regulators is key. Regulatory bodies need to lead in proactively enforcing by creating frameworks that allow innovation to flourish without undue risks. For example, regulations in the tech industry might focus on data privacy, ensuring new technologies don’t compromise personal identifiable information. In turn, this fosters trust and encourages quicker and wider adoption.

Above all, online platforms must keep children safe by preventing them from viewing or accessing age-restricted content or environments. Safety technology providers can assist platforms in the effective implementation of these tools to ensure compliance and help safeguard children and society online.

Read more:
Balancing innovation and regulatory compliance for online child safety 

E-commerce continues to grow in popularity with over 20% of purchases expected to occur online in 2024. This growth can be attributed to various factors, including online retailers now selling a variety of products and customers increasingly looking for discounts, greater choice and convenience.

In recent years consumers have increasingly used online retailers to purchase items including alcohol, tobacco, vaping and bladed items. As the sale of these age-restricted products and services becomes more common online, e-commerce platforms must be aware of their responsibilities, ensuring that only individuals of the correct age can purchase them.

Legislating E-commerce

E-commerce legislation is complex and challenging to navigate. Rules differ according to jurisdictions, cultural norms, geographic markets and regulatory frameworks. Requirements also vary from country to country in trading blocs, such as the EU and regionally such as state-to-state in the US.

Certain products such as alcohol, may be banned in some jurisdictions or regions with strict religious or cultural rules. Age requirements to purchase identical products may also vary between countries. For example, the US has a higher age requirement to purchase alcohol than European countries including the UK, France and Germany. There is no unified global approach, with individual countries and regions having a variety of legal frameworks in terms of the age assurance required.

Therefore, e-commerce businesses needing to implement age assurance solutions should look to both international standards and local legislation to guide their implementation. This is particularly important if they plan to sell age-restricted products internationally. Therefore, to comply with regulatory requirements, business leaders should understand and address the challenges surrounding age assurance. This prevents legal repercussions and makes their platforms safer.

Challenges

User Experience

As many consumers shop online for ease, age assurance processes must be robust yet frictionless, so as not to obstruct the user journey. If customers encounter difficulties during the purchasing process, they are likely to bounce to other (often less reputable) sites, resulting in a potential loss of customers. Therefore, the technology used on e-commerce sites must be easy to use and feel effortless for the customer.

Cross Border Challenges

In most cases, vendors sell to the same country and jurisdiction. This is particularly true for small and medium-sized businesses that are unlikely to sell in large quantities to warrant long-distance postage. However, international vendors must work case-by-case to ensure they align with international legislation.

Required Technology

When it comes to age assurance, there is no one-size-fits-all approach. Therefore, with age verification and age estimation offerings, optionality is key. The market requires flexible and inclusive options that address individual preferences while mitigating the risks associated with bias and exclusion. Methods include email address and facial age estimation, as well as verification methods including government-issued ID, credit card, mobile phone number and name & address checks.

When the site requests a customer to undergo an age estimation or verification process, platforms must communicate why their data is needed. This helps build confidence in the platform as consumers understand how their data is being handled and what it is used for.

Collaborating with Industry Partners

Online vendors have a duty to operate in a compliant way, while regulators are responsible for ensuring there is robust and equitable enforcement of related legislative requirements. If these regulations and processes are applied properly, underage users should not be able to ‘shop around’ to find the least stringent processes. Moreover, vendors should collaborate with a trusted and certified technology provider to ensure that their age assurance processes are both frictionless and compliant.

Safety Through Age Assurance

For online vendors, robust, frictionless age assurance solutions can improve safety, protect their reputation while ensuring underage consumers do not have access to age-restricted products. Those giving access to minors risk reputational damage or even regulatory repercussions.

Read more:
Trading Digitally: Understanding The Regulatory Requirements For Selling Age Restricted Products Online

The proliferation of artificial intelligence (AI) has been rapid, with 9 in 10 leading businesses (across 112 countries) leveraging some form of the technology within their operations.

New use cases are constantly emerging, with businesses across diverse sectors worldwide benefitting from enhanced analytics, productivity and a reduction in manual tasks.

In a rapidly evolving environment businesses want to implement AI to avoid falling behind and to remain competitive. As with any innovation, companies are interested in how they should implement AI while remaining compliant with upcoming regulations. As the landscape is currently unregulated, this is particularly important. The decisions companies make today will massively impact their future operations.

Rules around AI will likely relate to its general development and implementation. But these should also focus on specific use cases. This will be particularly relevant to content moderation, the practice of removing illegal, irrelevant or harmful material from online platforms.

We are already seeing an emergence of AI-generated images, many of which are age restricted or even illegal. AI-generated deepfakes have increased by 780% across Europe in the past year, with celebrities, politicians and law enforcement all highlighting the consequences of harmful AI-generated content.  Meanwhile, it is still unclear how legislators will cover content moderation in upcoming regulations. These changes will likely be significant, and something businesses will need to be aware of. Therefore, they will need to put the tools in place to identify and remove illegal content, including material that has been created using AI, and work with like minded businesses to tackle the issue.

Emerging AI Regulation

With the proliferation of AI, we can expect legislators across the globe to announce how they plan to regulate the technology in the coming months. The EU’s proposed Artificial Intelligence Act is a major step towards this as the first piece of draft regulation from a major legislative body and is expected to set the global standard for legislation in this area. The law will assign AI use into three categories: That which creates unacceptable risk, that which leads to high-risk. And other uses which are largely left unregulated.

Although the proposed AI Act is breaking new ground, it is expected to set the standard for these legislations worldwide. To help EU businesses prepare for implementation, the regulator has set up a tool for businesses to provide insights on regulatory obligations.

This tool defines who is subject to the legislation, based on a definition of AI and the role of the entity. Risk in this instance is related to how AI is used and the role of the organisation. As a consequence, companies can learn if they are subject to obligations if and when the regulation is approved. Currently, the exact wording and implementation of the act are still subject to debate.

For EU businesses, this legislation should be closely monitored to understand its impact. When operating in other jurisdictions, businesses should also monitor how they are impacted. While it is not the finished deal it will likely inform future legislation not just across the EU, but also the UK, US and beyond.

Unlike the internet, where regulatory action is still being debated decades on, we can expect AI legislation to be far swifter. Regulators, the business community and society in general have learnt lessons from the internet, which was left largely unregulated to allow for innovation and businesses to flourish. However, the online safety debate continues, with little solid guidance from regulators.

A universal standard of thinking will likely be established, building on the reality of AI implementation, the knowledge of third parties and regulatory necessities.

Preparing Your Business?

Businesses must keep an eye on key documents, launch timings and implications of upcoming AI legislation. Moreover, they must assess how they are deploying or expect to deploy AI in their organisations to understand their compliance obligations. As with any regulatory consideration, compliance likely comes with a monetary cost for businesses which will need to be factored in. For high-risk organisations, especially small or medium-sized enterprises, outsourcing these solutions to experts may be the most cost-effective, instead of developing and implementing in-house technology. Outsourcing solutions from vendors limits the need for in-house expertise and ensures that AI is used in an explainable way, crucial when under regulatory scrutiny.

For businesses impacted by regulations, a thorough understanding is essential, as is staying informed on relevant regulations and guidelines in every market they operate. For some business leaders, this task may seem daunting. However, there are simple steps to keep up.

Firstly, business leaders should engage with their peers, both in their industry and the wider ecosystem to keep up with new developments,  sharing best practices in working groups and industry forums. As subject matter experts, these platforms can help to address real issues relating to the implementation of regulation.

Impacted businesses should also perform regular audits and risk assessments to understand AI systems to understand compliance and risk. Before this process, businesses should maintain documentation of policies, procedures and decision-making processes. These documents can be used as evidence of compliance or to provide transparency to regulators and partners. To give a balanced view, these risk assessments can be conducted by third parties with broader experience.

Leaders must also put training and development in place at all levels for employees involved in AI development and deployment. Robust training will help employees to understand their responsibilities. This will ensure that these individuals understand their responsibilities regarding compliance and ethical AI use. By establishing a groundwork through training, businesses can implement continuous improvement practices, proactively addressing emerging challenges as they arise. Therefore, AI governance will be improved based on feedback, lessons learned and emerging best practices.

These practices will likely differ depending on industry, company size and function. In content moderation, however, it means aligning with like minded businesses to roll out a solution which can identify age restricted and illegal AI-generated content so that action can be taken accordingly.

Read more:
Business Best Practice: Preparing for AI Regulation

Businesses have historically lacked clarity from regulators regarding the compliant methods of age assurance available to them.

While some brands have had to devote significant time and resources to understand and implement the solutions that best ensure child safety on their platforms, others have taken a more passive approach, opting to wait for additional guidance from the regulators instead of taking proactive measures.

The recent comprehensive update to the UK’s Information Commissioner’s Opinion on age assurance for the Children’s Code is therefore a welcome development. It brings much needed clarity to businesses in relation to the implementation of the Children’s Code, by outlining a risk- and standards-based approach to age assurance. Furthermore, given the ICO’s broad remit focusing on overseeing and enforcing data protection laws in the UK, it helps businesses assess their compliance with obligations under the UK GDPR, as well as with wider regulatory frameworks.

Alignment on age assurance requirements with the critically important Online Safety Act 2023 shows further joined up thinking that will help protect our young people online. This Opinion drives a collaborative approach between regulators, alleviating contradictory requirements.

A Timely Announcement

The developments are particularly important in an environment where Ofcom has revealed that 75% of parents now worry about their children seeing age-inappropriate online content. And where access to smartphones has led to children becoming the biggest perpetrators of sexual abuse against children.

Businesses must know they need to do more when it comes to assessing the age of their users. Enforcing regulation and applied guidance from the ICO will propel this matter to the higher echelons of decision making, ensuring the adoption of age assurance tools more quickly and effectively.

Fit For Purpose Solutions

Naturally businesses will look for solutions that offer minimal friction for their users. The guidance around implementing readily accessible, privacy-preserving forms of age estimation such as email address are therefore pleasing. Not only will the majority of individuals have an email address and be willing to use it for the purposes of age estimation, this will complement biometric-driven age assurance including face and voice.

Verification methods based on ‘hard identifiers’ such as an ID scan are also viable, but can exclude or indirectly discriminate against people who lack the necessary documents or information, such as credit history or passports. Where possible, businesses should also consider offering a choice of age assurance methods, appropriate to the needs of their service and users.

A View To The Future

The ICO is being progressive in its approach by referencing techniques such as email address, which adds little or no friction to a user journey, making it both practical and commercially viable for platforms to adopt.

The note from the ICO stating that it will continue its engagement on international standards for age assurance technologies via the International Organisation for Standardisation and IEEE shows that this is a long-term commitment. Something that will be welcomed throughout the industry.

When choosing an age assurance solution, businesses must ensure they partner with an organisation that is not only able to implement the tools they require but one which is genuinely passionate about helping to build a safer online environment.

With it becoming increasingly easy for our young people to access age-restricted content, it has never been more important for regulators to implement and enforce legislation that will safeguard children and society online. Businesses must react quickly and ensure they have the tools and systems in place to ensure they are also doing their part.

Read more:
Providing Business Clarity: How updated Guidance from the ICO on Age Assurance for the Children’s Code will Improve Online Child Safety