Written November 21, 2023 by Jessica Downes, Esq., Staff Attorney, CHILD USA
AI-Generated CSAM: The Call for Proactive Action
Imagine your child comes home from school, hysterically crying, and tells you explicit photos of them are being shared online. To your shock, your child also tells you these photos are fake and were manufactured online by one of their classmates. This is the new reality of uncensored and unregulated generative artificial intelligence (“generative AI”). This exact scenario occurred at Westfield High School when a student used an online tool powered by generative AI to make pornographic images of female classmates and then shared these images with other boys at the school.[1] The victims of these images were devasted and many deleted their social media accounts in fear that more of their photos would be manipulated in such a way.[2]
For those unfamiliar with this new pop culture buzzword, AI is “the ability of a digital computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.”[3] Unlike previous AI which was designed to recognize patterns and make predictions based on data, generative AI is designed to create entirely new content. A commonly known software that uses generative AI is ChatGPT. While several studies have shown generative AI can have a positive impact on our society, such as self-driving cars and more personalized healthcare, generative AI has some dangerous pitfalls. AI-generated child sexual abuse material (“CSAM”) is one of them.
According to the 2023 Global Threat Assessment produced by WeProtect Global Alliance, an independent organization dedicated to combating online child sexual abuse, there is currently no evidence that child safety has been integrated into the design and implementation of generative AI services.[4] Though AI-generated CSAM accounts for less than 1% of all CSAM online, research by Thorn, an organization that builds technology to combat child sexual abuse, and Stanford University, shows the number of AI-generated CSAM has increased consistently since August 2022.[5] In a study conducted by the Internet Watch Foundation (“IWF”), IWF analysts found an online “manual” that teaches perpetrators how to refine prompts using generative AI to make more realistic CSAM.[6] Not only can generative AI assist with the creation of CSAM, but it can also serve as a powerful grooming tool by allowing perpetrators to elicit scripts on how to sexually extort children.[7]
Though many bills are circulating in Congress that address the harms of social media on children, including the EARN IT Act and The Kids Online Safety Act, there are currently no proposals aimed at addressing the precise harm that generative AI poses to children. However, it is notable that if passed the EARN IT Act would create a cause of action against Internet Service Providers (“ISPs”) regarding the “advertisement, promotion, presentation, distribution, or solicitation of child sexual abuse material,” which would provide survivors with an avenue to keep ISPs accountable in monitoring their platforms for CSAM.[8] Congress cannot remain ten years behind while emerging technologies operate unregulated, and scores of children are forced to endure their damaging effects.
On September 5, 2023, the National Association of Attorneys General, a bipartisan coalition of 54 state and territorial attorneys, issued a letter urging Congress to study how generative AI is being used to create CSAM and the effects such material has on children.[9] In May 2023, the former CEO of the artificial intelligence company that created ChatGPT, Sam Altman, stated at a Senate hearing that government intervention is imperative in mitigating the risks of generative AI.[10] During Altman’s testimony, he proposed the creation of a U.S. or global agency that would have the power to license AI systems and the authority to revoke these licenses in the event AI companies do not comply with safety standards.[11] While it is important to study the effects of generative AI, action must be taken now, rather than down the road, to prevent catastrophic harm to our children. Congress should create a right of action for survivors of generative AI CSAM to sue Generative AI companies for creating and producing software that is used to manufacture CSAM. Further, it should be a federal crime for someone to create and/or distribute guides on how to create generative AI CSAM.
Will Congress allow AI-generated CSAM to flood the internet before it acts? Or will it heed the sound advice of Altman, and others like him, and create a proactive, child-centered policy that prevents would-be perpetrators from weaponizing this technology and makes generative AI companies accountable for creating a harmful product? The safety and well-being of our nation’s children demand action now.
[1] See Julie Jargon, Fake Nudes of Real Students Cause an Uproar at a New Jersey High School, The Wall Street Journal, (Nov. 2, 2023)https://www.wsj.com/tech/fake-nudes-of-real-students-cause-an-uproar-at-a-new-jersey-high-school-df10f1bb?reflink=integratedwebview_share.
[2] Id.
[3] B.J. Copeland, Artificial Intelligence, Britannica (Nov. 10, 2023),https://www.britannica.com/technology/artificial-intelligence.
[4] WeProtect Global Alliance, 2023 Global Threat Assessment, 21 (2023), https://www.weprotect.org/wp-content/uploads/Global-Threat-Assessment-2023-English.pdf.
[5] David Thiel, et al., Generative ML and CSAM: Implications and Mitigations, 2 (2023), https://stacks.stanford.edu/file/druid:jv206yg3793/20230624-sio-cg-csam-report.pdf.
[6] Internet Watch Foundation, Prime Minister Must Act on Threat of AI as IWF ‘Sounds Alarm’ on First Confirmed AI-Generated Images of Child Sexual Abuse, (Jul 23, 2023), https://www.iwf.org.uk/news-media/news/prime-minister-must-act-on-threat-of-ai-as-iwf-sounds-alarm-on-first-confirmed-ai-generated-images-of-child-sexual-abuse/.
[7] David Thiel, et al., Supra 8.
[8] The EARN IT Act, S. 1207, 118th Cong. §5(6)(B)-(C) (2023).
[9] See NAAG, Attorney General, 54 Attorneys General Call on Congress to Study AI and its Harmful Effects on Children, (Sept. 5, 2023), https://www.naag.org/press-releases/54-attorneys-general-call-on-congress-to-study-ai-and-its-harmful-effects-on-children/.
[10] Matt O’Brien, ChaptGPT Chief Says Artificial Intelligence Should be Regulated US or Global Agency, AP News, (May 16, 2023, 5:53 PM), https://apnews.com/article/chatgpt-openai-ceo-sam-altman-congress-73ff96c6571f38ad5fd68b3072722790 .
[11] Id.