A recent study found that people are more likely to lie when using artificial intelligence (AI), suggesting a worrying trend in the way we interact with machines. The research involved over 8,000 participants across 13 studies and revealed that around 85% of people engaged in dishonesty when giving instructions to AI. This behavior was found to be more prevalent when individuals delegated tasks or gave prompts to AI systems, creating an environment where they could distance themselves from the consequences of their actions. The study’s authors warn that this trend has significant implications for high-risk areas such as finance, healthcare, and education, highlighting the need for further technical safeguards and regulatory frameworks to prevent the misuse of AI in these sectors.
Source: https://www.independent.co.uk/news/uk/home-news/ai-artificial-intelligence-lying-research-b2829104.html