In fact, it is hard to imagine what kind of difficulties people will encounter in the future for a technology that is in the limelight. Artificial intelligence and related technologies have been sought after for a long time, and common sense infers that its peak from prosperity to decline is coming soon. I had this hunch in the past 1.5 years, and Gartner's "hype cycle" in 2016 years confirmed this view.
Overall, AI? The momentum of "disillusionment" may have begun to appear. As a subjective inference, the time of disillusionment cannot be measured by scientific methods. However, whether you believe it or not, the popularity of AI is bound to rebound. However, the disillusionment of AI is not a bad thing to some extent, because in many cases, the content of hype far exceeds the current level of technological development.
But I'm not worried about this, because I'm more interested in long-term planning, including: can the outstanding achievements of the past few years be sustained? Can we continue to follow the completely distributed innovation model? Still counterproductive, AI will eventually usher in a cold winter?
I think the main reason why the development of AI will slow down is not the technology itself or the ability to create new solutions, but the public's enthusiasm for the implementation of these solutions. Volkswagen is very much looking forward to the application of level 5 autopilot technology, and is also very interested in the unmanned driving of aircraft, but at present the cockpit is still occupied by humans.
The confidence that the public has built up for many years may be destroyed just because of an accidental major accident. We can foresee that there will be more than one AI accident in the future. Although the accident rate of machine operation may be lower than that of manual operation, AI will still get a miserable end because of these "accidents".