We broke down the AI technologies of 2019 and shamelessly compared them with the forecast of 2017.
First, what is Gartner Hype Cycle? This is a kind of technology maturity cycle, or rather, the transition from the hype stage to its productive use. Now there will be a schedule with translation, so that clearly . Below are the explanations.
First stage. anger. Running. Technology appears, it is discussed first by enlightened nerds, and then by a fanatical public; excitement is gradually increasing.
Second stage. bargain. High expectations peak. At some point, everyone is already talking about technology, trying to implement it, and the most savvy - to sell at exorbitant prices.
Third stage. depression Decline of interest. The technology is being actively implemented and often fails due to flaws and limitations. "Bullshit is everything!" - comes here and there. The excitement drops sharply (the price tag, often, too).
Fourth stage. negation Work on bugs. Technology is improving, problems are being solved. Gradually, companies are carefully trying to implement technology and, hooray, everything is working out fine.
Fifth stage. Adoption Productive work. The technology is gaining its well-deserved place in the market and is quietly working, developing, and enjoying it.
What's in trend?
Returning to the 2019 Hype Cycle. Gartner in September, a report on which artificial intelligence technologies are at what stage and when they will start to work productively. Graph below, comments below graph.
With a large margin and already at the stage of "Productive work" are the technologies "Speech recognition" and "Acceleration of processes using the GPU." This means that they must be quickly applied, because they already provide a competitive advantage to their owners.
Automatic machine learning (AutoML) and chatbots are at the height of the hype right now. That is, everyone is talking about them, many are implementing them, but it will take from 2 to 5 conditionally to bring the technologies to the desired condition.
The cars we are used to are now also more than trendy. Autonomous vehicle technology is almost hitting rock bottom. In this case, this is good, because productive work lies ahead. However, Gartner estimates that it will take at least 10 years to develop and adapt.
Where are the once-hype drones and virtual reality now? Everything is in place - Gartner included drones in the field of Edge AI (categories bordering on AI), and virtual reality became part of Augmented intelligence (extended intelligence). Both topics, by the way, are now at the launch stage and have a positive outlook: 2-5 years before productive work on the market.
Prospects
Of the promising features: Robotic process automation software - it sounds scary, but in fact it is when the robot replaces routine actions. Nightmare of low-skilled personnel; however Harvard Business Review says there will be no layoffs, but productivity will increase. Eat believe. The technology will pass the peak of unpopularity and general contempt in 2 years, and then spread everywhere.
Of the technologies that evangelists and information gypsies of all stripes will talk about in large numbers only in the future, “neuromorphic equipment” was of particular interest. These are electrical devices (chips) that natural biological structures of our nervous system in terms of energy efficiency. To put it very simply, this is about super-productivity due to the division of labor (asynchronous updating of neurons). Giants such as IBM and Intel are already in full swing creating neuromorphic chips. But John Connor's army has time to prepare for the day of judgment - Gartner took as many as 10 years to mature the technology.
About digital ethics (Digital Ethics), which is typical, they talk a lot, but they are in no hurry to implement. The direction is singled out in a separate category of AI spheres: it means that it would be necessary to consolidate some ethical principles, norms and standards for data collection, implementation of AI in life, in general, so that it would be like . In the end, peep at Asimov.
Vs. 2017 2019
It's funny, but in 2017 everything was , even a separate AI hype cycle was not allotted: AI technologies were in the locomotive of emerging technologies (Emerging Technologies) along with blockchain and augmented reality.
Machine learning and deep learning in 2017 were on the hype Olympus, and in 2019 they continued their path towards decline, that is .
By the way, drones went from peak to decline for a year, and in 2019 they went back towards the approach to the peak. And it happens, yes.
In 2019, 8 new technologies entered the cycle. Among them are cloud services AI (Cloud Services), Marketplaces AI (Marketplaces), Quantum computing with AI (Quantum Computing). In general, well-known (in narrow circles) tools that are starting to be put on AI rails.
Source: habr.com