Hawley launched his expansive investigation last month after Meta’s AI chatbot was exposed for sending creepy and sensual messages to children.
Meta is beginning to turn over documents for Senator Josh Hawley’s (R., Mo.) sweeping investigation into its artificial intelligence chatbots.
Meta handed over a batch of documents earlier this week after an unexpected transmission issue delayed the process, a Meta official told NR. Meta plans on producing more documents in the near future in order to fully comply with Hawley’s wide-ranging request for materials.
Hawley’s deadline for the documents passed last week, before Meta produced the first batch and began to comply with his request. He launched his expansive investigation last month after Meta’s AI chatbot was exposed for sending creepy and sensual messages to children.
“The Senate Judiciary Committee Subcommittee on Crime and Counterterrorism, which I chair, will commence an investigation into whether Meta’s generative-AI products enable exploitation, deception, or other criminal harms to children, and whether Meta misled the public or regulators about its safeguards,” Hawley wrote to Meta CEO Mark Zuckerberg.
“We intend to learn who approved these policies, how long they were in effect, and what Meta has done to stop this conduct going forward.”
He asked for Meta to produce a trove of documents on its AI risk standards and their implementation. His request included documents on the enforcement of the AI risk standards and how Meta handled incident reports.
Once the conduct of Meta’s AI became known, the company changed its guidelines to limit teen access to AI characters and prevent the chatbots from responding to sensitive topics.
Hawley is part of a group of GOP senators who raised concerns earlier this month about Meta’s transparency when it comes to its efforts to protect children from harmful content. Those concerns were amplified at a hearing on September 9 when two whistleblowers came forward with allegations Meta buried child safety research and ignored evidence of negative user experiences.
The whistleblower allegations were primarily oriented around the psychological damage to children subject to bullying and forms of sexual misconduct in Meta’s virtual reality worlds. Meta dismissed the whistleblower allegations as falsehoods based on cherry-picked internal documents.
A week later, Hawley held a subcommittee hearing on AI chatbots encouraging children to commit suicide and engage in other destructive patterns of behavior. Parents of deceased and mentally unwell children delivered explosive testimony on the ways the AI chatbots became close confidants to their children and pushed them into the depths of despair.
Open AI and Character AI, two companies whose products were accused of driving teens to suicide, have both made substantial upgrades to their safety policies and parental controls to prevent such instances from happening again. At the hearing, Hawley floated the possibility of compelling testimony from the CEOs of Meta, Open AI, and Character AI. It remains to be seen whether he will do so to continue his investigation.
More broadly, Hawley has warned about the potentially negative impacts of AI on the human experience if it is not harnessed properly. He believes AI could undermine American ideals if it is used to dismantle livelihoods and reduce American liberty instead of improving the lives of working people.