Abstract Previous research has shown that consumers respond differently to decisions made by humans versus algorithms. Many tasks, however, are not performed by humans anymore but entirely by algorithms. In fact, consumers increasingly encounter algorithm-controlled products, such as robotic vacuum cleaners or smart refrigerators, which are steered by different types of algorithms. Building on insights from computer science and consumer research on algorithm perception, this research investigates how consumers respond to different types of algorithms within these products. This research compares high-adaptivity algorithms, which can learn and adapt, versus low-adaptivity algorithms, which are entirely pre-programmed, and explore their impact on consumers' product preferences. Six empirical studies show that, in general, consumers prefer products with high-adaptivity algorithms. However, this preference depends on the desired level of product outcome range—the number of solutions a product is expected to provide within a task or across tasks. The findings also demonstrate that perceived algorithm creativity and predictability drive the observed effects. This research highlights the distinctive role of algorithm types in the perception of consumer goods and reveals the consequences of unveiling the mind of the machine to consumers.

People turn to generative artificial intelligence (AI) to make ideation less effortful. Turning a vague idea into a mature concept is hard work, and offloading it to AI is tempting. However, our research shows that this strategy can backfire; the lower-effort form of AI support often produced less creative ideas. In an online experiment, 276 people refined ideas for an innovation challenge working alone or with AI-generated text or images. Image-based generative AI had a double-edged effect; it significantly reduced effort compared with working alone or using text-based support. However, ideas from image-based support were also 18% less creative than those from the more effortful text-based support. The reason lies in what each format leaves for the human to do. Images specify all details of an idea, leaving less creative room for humans, but text leaves gaps that humans can complete with their imagination—a process that is effortful yet stimulates creativity. Crucially, the effect also depends on how far the idea is developed. Text-based input helps most when the idea is already well developed because humans can then make the most of it. The major takeaway is do not use AI just to save effort but use it where the effort pays off.

Following trends on social media has become increasingly popular. But what is the best way to do so? Should brands and other creators copy the trend as closely as possible, or should they put a more unique spin on it? To answer this question, the authors develop a multimodal, unsupervised video analytics tool (MUVID) to quantify the typicality of over 85,000 TikTok dance videos. Results indicate that more atypical videos (i.e., more differentiated from the trend) generate more engagement. Consistent with the notion that atypicality drives engagement, this relationship is amplified when atypicality is easier to observe (i.e., when audiences have seen more trend videos). Follow-up experiments, including a content-creator field experiment, manipulate atypicality and confirm its causal impact. The findings provide practical guidance on how to create more impactful content, shed light on effective trend-following, and offer a tool (available through an app) that researchers and practitioners can use to quantify typicality and analyze short videos more generally.