What the Online Safety Act 2023 Regulates and Who Is In Scope
The Online Safety Act 2023 introduces new obligations for online platforms and content creators in the UK. The Act applies to providers of online platforms, including social media networks, messaging services, and content hosting sites. These entities must take steps to protect users from harmful content, including illegal content, harassment, and content that could cause psychological harm. The Act places a duty on platforms to monitor and respond to reports of harmful content, and to take proactive measures to reduce risks.
The Act also applies to providers of AI-generated content. This includes platforms that host or generate content using artificial intelligence, such as deepfakes, AI-generated images or text, or automated content creation tools. The focus is on ensuring these platforms implement appropriate safeguards to prevent misuse or harm caused by such content. The Act makes it clear that AI-generated content must be identified or marked where required, particularly when it mimics real people or existing content.
The scope of the Act covers any platform that makes content available to users in the UK. This includes global platforms that have a significant presence or user base in the UK. The Act applies to both hosted content and content generated through AI tools. For example, a social media platform that allows users to upload AI-generated images or posts must ensure these are identified or marked appropriately. Similarly, a content creator who uses AI tools to generate text or images must ensure these are properly identified if they are later published or hosted on a platform.
The Act also places responsibilities on AI developers and providers. Where AI systems are used to generate or host content, these providers must ensure that the systems are designed to prevent or reduce harm. This includes considering the potential for misuse or misrepresentation. For example, an AI tool that generates deepfake videos must have safeguards to prevent the creation of misleading or harmful content. The Act makes it clear that these responsibilities apply to any AI system that is used to create or host content, regardless of whether the AI is hosted on a platform or used by an individual.
The Act introduces a framework for addressing harmful content through a combination of proactive measures and reactive responses. Platforms must have systems in place to identify and remove harmful content quickly. They must also have processes for addressing user reports of harmful content. For example, a messaging platform must have a way to identify and respond to reports of harassment or threats. The Act requires platforms to publish annual reports on their efforts to tackle harmful content, including data on the number of reports received and actions taken.
The Act also places duties on platforms to ensure that AI-generated content is identified or marked. This is particularly important for content that could mislead or deceive users. For example, an AI-generated news article must be clearly identified as such. The Act makes it clear that platforms must take steps to ensure that users understand when they are viewing AI-generated content. This includes providing clear information about the source or nature of the content.
The Act applies to any platform or AI developer that makes content available to users in the UK. This includes platforms that are based outside the UK but have users in the UK. The Act also applies to any AI tool or system that is used to generate or host content. For example, a UK-based company that uses AI to generate marketing content must ensure that this content is identified or marked appropriately. The Act makes it clear that these responsibilities apply to any entity that is involved in the creation or hosting of AI-generated content.
The Act introduces a range of duties that apply to platforms and AI developers. These include obligations to monitor content, respond to reports, and take steps to reduce harm. The Act also requires platforms to have processes in place for addressing user concerns. For example, a platform must have a way to respond to user complaints about AI-generated content. The Act makes it clear that these duties apply to any platform or AI developer that makes content available to users in the UK.
The Act places responsibilities on platforms to ensure that AI-generated content is identified or marked. This is particularly important for content that could mislead or deceive users. For example, an AI-generated image that mimics a real person must be clearly identified. The Act makes it clear that platforms must take steps to ensure that users understand when they are viewing AI-generated content. This includes providing clear information about the source or nature of the content.
The Act also makes it clear that AI developers must take steps to prevent or reduce harm. This includes considering the potential for misuse or misrepresentation. For example, an AI tool that generates deepfake videos must have safeguards to prevent the creation of misleading or harmful content. The Act makes it clear that these responsibilities apply to any AI system that is used to create or host content, regardless of whether the AI is hosted on a platform or used by an individual.
The Act introduces a framework for addressing harmful content through a combination of proactive measures and reactive responses. Platforms must have systems in place to identify and remove harmful content quickly. They must also have processes for addressing user reports of harmful content. For example, a messaging platform must have a way to identify and respond to reports of harassment or threats. The Act requires platforms to publish annual reports on their efforts to tackle harmful content, including data on the number of reports received and actions taken.
The Act makes it clear that these duties apply to any platform or AI developer that makes content available to users in the UK. The Act also applies to any AI tool or system that is used to generate or host content. The focus is on ensuring that platforms and developers take steps to prevent or reduce harm caused by AI-generated content. This includes identifying or marking such content, and ensuring that users understand when they are viewing AI-generated material.
