New Analysis of Generative AI Chatbots Shows Signs of Improvement and Risks for Kids
Common Sense Media's reviews of Claude, Gemini, and Perplexity caution that real-time search features may mislead young users into thinking responses are fact-based
SAN FRANCISCO, Aug. 27, 2024—Today, Common Sense Media published a new set of reviews that assess the ethical use, transparency, safety, and potential risks of popular AI chatbots, including Google's Gemini and Anthropic's Claude, which received risk level ratings of "low" and "minimal," respectively. It also reviewed Perplexity, which received a "high" risk rating.
The reviews show that while some chatbots are doing a better job of addressing obviously harmful stereotypes and blatant misinformation, they may be creating a false sense of security and credibility, especially for young and impressionable minds. The reviews caution users about AI chatbots that use real-time search like in Perplexity or "double-check" features like in Gemini because, while these can enhance transparency, Common Sense findings show the output is not always accurate.
"Generative AI chatbots are touted as tools that save users' time, but in actuality they should simply reallocate it," said Tracy Pizzo-Frey, senior advisor on AI to Common Sense Media. "Our robust analysis of several popular chatbots makes it clear that users—kids and teens especially—should engage their critical thinking skills and spend more time verifying AI's results, rather than risk overreliance and accept the initial results as fact."
On the upside, some companies are taking steps to prevent negative experiences with chatbots for kids and teens. Google's Gemini, which is currently the only major platform to create a specific experience for teens, includes additional CSAM (child sexual abuse material) safeguards, content policies, and in-product experiences that help teens understand the limitations of generative AI.
"The development of AI chatbots and new features is outpacing consumers' ability to understand them, which could compromise the safety and privacy of children and teens," said James P. Steyer, founder and CEO of Common Sense Media. "We hope these AI reviews will keep users informed and encourage industry transparency so that more developers can continue to innovate in a responsible and ethical way."
Developed with leading AI experts, Common Sense Media's AI reviews are designed to be "nutrition labels" for the public. They identify which tools are using best practices—and which may compromise human rights or data privacy, or perpetuate misinformation and unfair bias.
For more information, please visit commonsense.org/ai.
About Common Sense:
Common Sense is dedicated to improving the lives of kids and families by providing the trustworthy information, education, and independent voice they need to thrive. Our ratings, research, and resources reach approximately 150 million users worldwide and 1.2 million educators per year. Learn more at www.commonsense.org.