Some experts are worried that many of these new technologies are not able to discern accurate and truthful information from misinformation or age-appropriate advice from age-inappropriate advice and the potential risk this poses for the millions of young children interacting with these new technologies.
Concerns relate to the way AI products are being designed and presented, with increasingly human-like characteristics, appearances, and speech patterns that are designed to deliver convincing, authentic-sounding output. This, some experts feel, may lead children to believe that what is being delivered by ChatGPT and AI chatbots is completely true and accurate, when in fact, they could be consuming content that is taken from unreliable, ill-informed, or even completely fictitious sources.
Three characteristics of generative AI increase the issue and potential for harm arising from mis- and disinformation. One is the ease and incredible efficiency of content creation, and the second one is the well-polished and authoritative-sounding form of the output, whether ChatGPT played fast and loose with reality or was as accurate as truth itself. Third, generative AI has the ability to appear human, form emotional connections and become a trusted friend like a conventional search engine never could. In combination, these three characteristics may significantly increase the likelihood that the output produced by ChatGPT is taken for sound information.
Patricia Thaine, Forbes Technology Council
Thaine, CEO of Private AI and member of the Forbes Technology Council, has suggested a number of ways to improve ChatGPT and AI chatbot child safety, which include:
- Realtime Factchecking: Development of real-time fact-checking that enables AI to analyze incoming and outgoing data for accuracy by cross-referencing all content against a wide range of reliable databases and trusted news networks.
- Transparency Labels: The use of labels or tags to inform users of the nature of content delivered by ChatGPT and AI chatbots. For example, 'AI Generated Advice' or 'Unverified Information' tags could help users understand they are not talking to a human and that the information or content they are consuming may not be reliable or accurate.
- Parental Approval: This could include linking a child's account to an adult account, enabling parental control over what a child can access.
- Multi-Factor Age Verification: Improved age verification processes that include more levels of verification, such as facial recognition or legal document checks for underage users.
The concerns surrounding AI-generated interactions and content creation are not new, and as these technologies continue to grow, the need to develop more technology-based solutions to protect the young and vulnerable also grows and will hopefully not get left behind in the race for commercial success.