Artificial intelligence chatbots like ChatGPT are becoming increasingly common tools in our daily lives, but can we really trust them to provide accurate and safe health information? Consumer Reports, a renowned technology and privacy research organization, recently conducted a series of experiments to determine just that.
AI Accuracy Testing
Consumer Reports put five popular AI chatbots to the test to see if their advice on health and safety topics aligned with that of industry experts. The results were eye-opening.
Carbon Monoxide Detector Placement
When asked how many carbon monoxide detectors are needed in a home, CR recommended one on each level, outside each sleeping area, in the basement, and near—not inside—an attached garage. While most chatbots got this right, Google Gemini made a critical error by suggesting placing a CO alarm inside an attached garage, which can damage the device's components due to temperature fluctuations.
Filtering PFAS from Tap Water
CR advised using a water filtration system with the NSF/ANSI 53 certification and ensuring the manufacturer explicitly claims PFAS removal. However, AI responses, such as those from Microsoft Copilot, were lacking in crucial details, potentially leading consumers to purchase unsuitable filters.
Water Beads Safety
When asked about the safety of letting kids play with water beads, Consumer Reports strongly advised against it due to the high risk of ingestion-related injuries. In contrast, AI responses, like those from Meta AI, failed to emphasize the significant dangers associated with these toys.
Overall, while AI chatbots can be convenient, it's essential to remember that they are not infallible experts. Consumer Reports recommends using AI as a helpful tool but always double-checking important health and safety advice with reliable sources. After all, when it comes to your well-being, it's better to be safe than sorry.