As the phrase “Artificial Intelligence” is applied to cover an ever-widening range of computational practices, attracting both attention, research and funding to all sorts of private and public ventures, it has also been proposed as a solution to everything from criminality to environmental collapse. While the very real benefits afforded by e.g., automation and digital communication platforms are hard to deny, it is also difficult to believe in many of the utopian visions proposed by AI evangelists.  
The problem is, broadly, two-fold: AI Hype not only presents the current state of various AI projects as more advanced and fundamentally impressive than they are (many ‘AI’ applications rely heavily on Mechanical Turk and similar platforms for a large proportion of their work), but also overpromises on the near- and long-term capabilities of AI, to the point where it is portrayed as being able to tackle virtually every conceivable problem, if applied in the ‘correct’ way (what, exactly, that means, is often left unstated, underspecified, or is prima facie absurd). Meanwhile, because of this wide-spread hype, AI Solutionism leads to proposals to solve every actual existing problem with ‘AI’, which both leads to problems not being solved (due to the applied AI solution not being up to the task, or not yet existing), as well as a disproportionate lack of resources given to existing and proposed solutions not involving AI.
This situation poses a number of questions to the academic AI community; what is our role as AI researchers in countering (or driving) hype about AI, and what should it be in the future? How can and should we act within the current system to achieve the research and results we wish for? How can and should we act to change the current system into a better one? What are the appropriate laws and regulations to propose to policymakers, and at what levels should they be applied?