Markets 
Policymakers risk repeating mistakes they made trying to govern social media in their rush to protect kids from potential harms posed by AI chatbots, according to a new report by a Washington, D.C. think tank.
In seeking solutions to real or alleged harms to children on social media, policymakers turned to flawed policies — such as age verification and outright bans — that cut children off from social media’s benefits instead of actually targeting its risks, the report published by the Information Technology & Innovation Foundation (ITIF) maintained.
The report, written by policy analyst Alex Ambrose, contends that legislators are now recycling that playbook for artificial intelligence (AI) chatbots, or systems that process user inputs to respond with simulated conversation.
But the playbook doesn’t work for either AI or chatbots, it asserted.
At the same time, it continued, recent unfortunate events such as child suicides, violent shootings and radicalization of young people establish a clear need for effective chatbot regulations regarding children’s online safety.
But importing ineffective social media frameworks to chatbots fails to properly address children’s safety concerns while, at minimum, inconveniencing adult users, it added.
Focus on Targeted Safeguards
The report argues that the best ways for policymakers to address chatbot safety for children are to focus on transparency, targeted safeguards, and empowering parents.
“Focusing on transparency, targeted safeguards, and giving parents more control makes chatbots safer because these approaches can address documented harms, such as conversations delving into topics of extreme physical or mental harm, and features that enable unhealthy parasocial relationships with chatbots, while not overly restricting access to the technology,” Ambrose told TechNewsWorld.
Transparency lets parents know what the chatbot can do and what data it collects, explained Rohit Poduval, a senior software engineer at Amazon and a member of the Integrity Institute, a global think tank made up of tech workers with hands-on experience identifying and mitigating harms on social media and internet platforms.
“Targeted safeguards address specific harms without blocking legitimate use,” he told TechNewsWorld. “For example, you can prohibit a chatbot from simulating romantic relationships with minors without blocking homework help.”
“Parental controls let families set boundaries that match their child’s maturity,” he said. “These three together put control where it belongs — with informed parents, appropriate access, clear rules for platforms.”
“Of course, transparency and targeted safeguards are important to have in place with chatbots. However, it’s crucial that parents understand the risks they need to look out for to begin with,” added Jeffrey Nadrich, founder and managing attorney of Nadrich Accident Injury Lawyers in Los Angeles.
“In many instances, parents have no idea their child is having multi-hour conversations with a chatbot online,” he told TechNewsWorld.
Parental Controls Have Limits
Madeline Pendley, an attorney with Rafferty Domnick Cunningham & Yaffa in Palm Beach Gardens, Fla., pointed out that while transparency, targeted safeguards and empowering parents can certainly make chatbots safer, they are only effective when platforms actually build, maintain, and require the use of those tools.
For example, the child-flag system proposed in the report is a great tool that should be implemented by AI platforms. “However, as applied, it still puts a duty back on parents to affirmatively flag the device,” she told TechNewsWorld.
“In theory, this could be one of many required features that could prevent AI-based harm, but is, on its own, not sufficient,” she argued.
“This is true of many of the parental controls on web-based platforms,” she said. “They are easily circumvented, or parents do not know to apply them.”
“Empowering parents and encouraging digital literacy is certainly crucial,” she added, “but the obligation is not on the parents to ensure the platforms are safe. It’s on the platform developers.”
Too Much Weight on Parents
Transparency and targeted safeguards are absolutely the right priorities, agreed Whitney Ray Di Bona, an attorney and consumer safety advocate at Drugwatch, a website focused on dangerous prescription drugs and defective medical devices. “Where I part ways with this report is the weight it places on parental supervision as the primary solution,” she told TechNewsWorld.
“In my opinion, placing the burden on parents to protect their children from the harms of AI chatbots fundamentally misunderstands the reality most families are living in,” she argued.
“Many parents do not know their children are using these tools,” she continued. “Many do not fully understand what AI chatbots are or how they work. And many are simply not in a position to monitor their child’s technology use around the clock.”
“That is not a failure of parenting,” she said. “That is a reality that policymakers and companies need to account for.”
Parental controls alone are not enough, added Alex Zeig, CEO and founder of BlueCheck, an identity and age verification company in Austin, Texas. “A parent can lock down a home device and still have their kid watching unmoderated content on a school Chromebook,” he told TechNewsWorld. “That’s not a failure of parenting because parents can’t win a game that spans every device, every network, every platform their kid might encounter.”
“Here’s where I think age verification is important,” he continued. “If a platform knows it’s talking to a 12-year-old instead of a 32-year-old, it has to actually fix what it’s serving that child. Age verification doesn’t just help parents. It changes who’s accountable.”
Broad Restrictions Raise Privacy, Speech Concerns
The ITIF report also maintained that many bills to protect children from chatbot harms err by simply transplanting ineffective social media approaches, such as age verification, content restrictions, and blanket bans.
“These provisions are ineffective because, in their attempt to protect children from harms, these policies restrict both kids’ and adults’ access to information,” the ITIF’s Ambrose explained.
“They also take the decision out of the hands of parents who may be better able to decide what AI chatbots are appropriate for their children based on their own knowledge about their children’s maturity, level of adult supervision, and intended use,” she added.
The new ITIF report rightfully argues that broad bans, content restrictions, and age verification for chatbots are ineffective, counterproductive, and raise First Amendment issues, agreed David Inserra, a fellow for free expression and technology at the Cato Institute, a Washington, D.C. think tank.
“Mass age verification requirements for AI raise major privacy concerns that also chill speech,” he told TechNewsWorld. “As the ITIF report argues, requiring an ID, biometric, or other form of identifying information creates privacy risks by requiring AI companies to collect and store data about individuals using their services.”
“Content restrictions and broad bans also present more direct free-speech concerns,” he continued. “Requiring companies to keep kids safe from content that is violent, harassing, resulting in depression or anxiety, or otherwise dangerous to minors is content-based censorship that is constitutionally dubious, and for good reason. Americans disagree on what types of content are harmful or dangerous for kids.”
Developers Left Off the Hook
Among the strongest recommendations in the report are those around controls and education, according to Giselle Fuerte, founder and CEO of Being Human With AI. These include giving parents control over features such as sycophantic language and building digital literacy so children can apply their developing critical thinking skills to chatbots.
Parental controls can fill the gaps where kids aren’t yet cognitively ready to manage the risks the way adults can, added Fuerte, whose Spokane, Wash., company focuses on education about artificial intelligence ethics.
“But this framing lets developers off the hook,” she told TechNewsWorld. “It takes the controls of a wildly unpredictable text-prediction system — one with a ‘black box’ of reasoning so opaque that an entire field of AI researchers now exists to untangle it — and hands them to already overworked parents, asking them to be the ultimate gatekeepers of a system implicated in the deaths of children like Adam Raine. That’s a heavy burden to offload onto the consumer.”
“The report doesn’t go far enough in mandating design restrictions that would put responsibility on developers to ship products that are safe in the first place, rather than on parents to make sure their child doesn’t cut themselves on the jagged edge of a design defect,” she said.
With effective protections in place, chatbots and other AI-driven technology can be beneficial tools for children, added Jessica Bregant, an assistant professor of law and associated faculty in the department of psychology at the University of Houston.
“The most obvious benefit is fluency with a technology that seems likely to be a major part of work and life for them in the future,” she told TechNewsWorld.
“I am encouraged by new studies showing that working with AI chatbots may actually support learning rather than supplanting it, as so many have feared,” she added. “Children are especially flexible in their thinking and learning, and today’s children will probably develop ways of interacting with AI that we can’t yet imagine.”
“For better and worse, this technology is part of the culture, and empowering children to use it responsibly is going to be a challenge for everyone who is a parent or an educator,” she said.
Read the full article here
