Meta has announced its plans to collaborate with a few tech companies to establish a framework of principles to mitigate risks arising from the development of genAI.
The company is partnering with nonprofit organisations such as All Tech is Human and Thorn in a bid to curb the use of GenAI tools for child exploitation, even through product development processes.
Meta also entered into a similar partnership with IBM and other tech companies called the AI Alliance to encourage the development of responsible, open, and safe AI tools. The initiative even encouraged open-source products and generated educational resources that could benefit policymakers and the public.
Meta’s latest effort, however, largely focuses on preventing child exploitation. Consequently, the partnership will look to create guidelines for AI developers throughout the development, deployment, and maintenance phases of a product life cycle.
Some of the guidelines include responsible sourcing of datasets used to train AI models to avoid materials with child sexual abuse and exploitation. Meta is also looking at stress testing and content provenance solutions that will make it easy to identify AI-generated content.
Other measures include responsible hosting of first-party AI models, supporting developer ownership, and creating safeguards against malicious content and conduct by users.
Moreover, the collaboration will work to develop tech solutions to prevent malicious AI applications and eliminate exploitative content from major platforms.
The development comes when Meta has frequently been under fire by government bodies for instances of child exploitation, mental health issues, and rights violations on its platforms.
The move could be seen as necessary for the social media company to keep up with the increasingly sophisticated methodologies being used by bad actors in recent months.
