Contemplating possible future scenarios should be left to the field of philosophy, not to futurology – or so claims Luciano Floridi in this somewhat harsh but fair editor letter. Floridi examines so-called Artificial Intelligence (AI) winters and their impact on the development of AI. AI winters are periods where the hype for AI wanes and often the result of disillusionment when – inevitably – promises concerning AI applications fail to deliver.

The drawbacks of hype for AI are twofold: on the one hand, it makes people more skeptical towards useful applications, and on the other hand alarmism blinds people to the actual risks associated with AI applications. Floridi mentions these instances of alarmism during his discussion of hype, seemingly categorizing alarmism as an instance of hype rather than a different variant of an exaggerated claim.

Floridi is afraid that when the hype bubble bursts, ‘unreasonable disillusionment’ will follow. Instead of judging individual applications of AI on their merit, funders, journalists, and other stakeholders might disregard AI as a whole, ignoring potentially useful technologies. A similar argument applies to alarmism. Instead of worrying about AI as some future existential risk, Floridi argues, we should worry about current problems such as ‘increased pressure on individual and group privacy’, identity theft, and cyberconflicts. However, alarmism either leads stakeholders to focus on the wrong problems or to disregard the dangers of AI entirely.

Floridi provides a mundane but effective recommendation to prevent hype, namely to ask the following questions: what is the social acceptability of new technology? Are the necessary skills, datasets, infrastructure, and business models in place to make such a technology successful? Contemplating these issues might be less exciting than wild speculation, but much more constructive.