In recent years, major technology companies such as Microsoft, Meta, Google, and OpenAI have been focusing on developing generative Artificial Intelligence (AI) tools to combat child sexual abuse images (CSAM). These companies have committed to implementing security measures by design to address this issue.
According to child safety organization Thorn, over 104 million files suspected of containing CSAM were reported in the United States in 2023. The influx of AI-generated images poses significant risks to children and further burdens an already strained child safety ecosystem. To tackle this problem, Thorn and All Tech is Human, along with companies like Amazon, Meta, Microsoft, and Google, have joined forces to promote an initiative aimed at protecting minors from the misuse of AI.
The technology firms participating in this initiative have pledged to adhere to security by design principles. This involves implementing measures to prevent the creation of harmful content. Cybercriminals can leverage generative AI for various purposes, including making it difficult to identify victims and increasing demand for abusive material. By proactively addressing child safety risks in the development of AI models, these companies are working to prevent the dissemination of harmful content.
Some of the measures included in security by design principles involve studying training data to ensure it does not reproduce abusive content, implementing watermarks in AI-generated images, and evaluating models for child safety before distribution. Technology giants like Google have also developed tools to detect and remove CSAM material on their platforms. For example, they use hash matching technology and AI classifiers to identify potentially harmful content and collaborate with organizations like the National Center for Missing and Exploited Children (NCMEC) to report incidents.
By prioritizing child safety and investing in technological solutions, these companies are demonstrating their commitment to combating the misuse of generative AI for harmful purposes. Through collaborations and initiatives like the one described above
