Home > Electrical/Electronics > Computer (I.T) > Tech Giants Pledge To Eliminate AI-Generated CSAM

Tech Giants Pledge To Eliminate AI-Generated CSAM

Tech giants have promised to make sure that their AI training datasets are free of any child sexual abuse material (CSAM).

In collaboration with nonprofit Thorn and All Tech Is Human, the firms—Amazon, Anthropic, Civitai, Google, Meta, Metaphysic, Microsoft, Mistral AI, OpenAI, and Stability AI—have committed to a series of ‘safety by design’ principles aimed at stamping out AI-generated CSAM.

“In the same way that the internet accelerated offline and online sexual harms against children, misuse of generative AI presents a profound threat to child safety—with implications across child victimization, victim identification, abuse proliferation and more,” Thorn said in a statement.

“This misuse, and its associated downstream harm, is already occurring—within our very own communities. Yet we find ourselves in a rare moment—a window of opportunity—to still go down the right path with generative AI and ensure children are protected as the technology is built.”

The firms will, they say, source their training datasets responsibly, avoiding or at least mitigating training data with a known risk of containing CSAM, removing it from use and reporting any confirmed CSAM to the relevant authorities.

They’ll incorporate feedback loops and iterative stress-testing techniques into their development process, and use content provenance solutions to work out whether content is AI-generated. The firms are also committing to safeguarding their generative AI products and services from abusive content and conduct, and to hosting their models responsibly.

They will, they say, prevent their services from scaling access to harmful tools, and invest in research and future technologies to try and wipe out the use of generative AI for online child sexual abuse and exploitation.

“These mitigations complement our existing work to prevent the creation, dissemination and promotion of AI generated child sexual abuse and exploitation,” said Susan Jasper, Google’s VP of trust and safety solutions.

She said the company is integrating both hash-matching and child safety classifiers to remove CSAM, as well as other exploitative and illegal content, from its training datasets.

READ ALSO  Urgent Facebook Messenger Warning For Millions Users

Last December, an investigation by the Stanford Internet Observatory identified hundreds of known CSAM images in an open dataset of billions of images, known as LAION-5B, used to train popular AI text-to-image generation models. The image URLs were reported to the National Center for Missing and Exploited Children in the U.S. and the Canadian Centre for Child Protection, and are now being removed.

However, the National Center for Missing & Exploited Children recently said that it’s struggling to keep up with the rising number of CSAM reports it receives, up by 12% in 2023 to hit 36.2 million reports in total. Of these, 4,700 involved CSAM or other sexually exploitative content that was related to generative AI.

“Legislation and regulation is needed to ensure that GAI technology is not trained on child sexual exploitation content, is taught not to create such content, and that GAI platforms are required to detect, report, and remove attempts to create child sexual exploitation content and held responsible for creation of this content using their tools,” said the center.

Source: Forbes

Total Views: 212 ,
0
0

Leave a Reply

Your email address will not be published. Required fields are marked *