To design sustainable AI designers must be able to understand and think about complex technical, ecological, social, and economic systems and their interactions. Their reasoning and decisions need to be based on ethics and scientific facts. They must acknowledge different stakeholders’ social and cultural norms, practices and current and future needs. Unfortunately, designers’ thinking is prone to err, biases, and other psychological phenomena, which can negatively affect how they understand, reason, and make decision, and which can lead to unsustainable and unethical AI solutions. Thus, it is important to investigate errors in designers’ thinking. This study presents a cognitive scientific overview about some common errors when making arguments, inferring, and reasoning, when drawing analogies, or in situations where problems are complex, uncertain, challenging the status quo, or framed differently. Also, processing information, emotions and social and cultural aspects can be source of errors in thinking. Designers must become aware of the risk of errors in their own perceptions, thinking, and reasoning and to explain why, what, and how they design sustainable AI. This can lead to more ethical and sustainable solutions in AI design.