There is no doubt about it that the year 2020 will be a milestone for artificial intelligence with tech trends such as quantum AI, deep-fake celebrities, AI malware … etc.
Experts’ opinion is that in 2020 we can expect different AI systems to be moved from testing and experimental use, to applied and practical usage by corporations that are wealthy enough to deploy such systems and harness their power.
What is this AI transfer from testing to practical use means?
Before progressing further I would note that current experimental and applied AI is still far away from a “general AI with cognitive functions”. Instead, it is a large number of refined algorithms that process large quantities of data, notice patterns at incredible speeds well beyond human capabilities, and provide results for a specific task.
What this means is that, so far, different AI systems have been in the testing phase with large quantities of data being fed into them, the results being monitored and algorithms tweaked. This process is called “training an AI”, with the goal being a more accurate result. Those systems that are in the final phase of training will be moving to the applied phase.
Which challenges will be applied to Artificial Intelligence incur?
One of the more troubling artificial intelligence applications is the celebrity deep-fake which makes it difficult to discern real media from fake. Malicious use can lead to misinformation of the public, as well as swaying public opinions.
On the other hand, commercial deep-fakes can be used to “resurrect celebrities” and completely change the media and entertainment landscape.
Finally, deep-fakes renounce accountability for any publicly displayed video content, where the person in question can deny its own presence in the content and declare it a deep-fake, making it will be hard to discern the truth.
What goals have yet to be achieved with Artificial Intelligence?
Because this kind of technology takes a long time to develop, we can say that in the relatively short-term within the next five years what we can expect is for the AI platforms to become more energy-efficient, with the integration of quantum neural networks and progress into natural language processing.
Arguably, the main issue with current artificial intelligence systems is that they require an enormous amount of data to be trained in the first place, meaning new techniques that provide similar results within the margin of error with significantly fewer data will be in the focus.
Conclusion
Up to now, companies that deploy predictive inventory management and inventory flow, data-driven brand decisions regarding sales, and other AI techniques will hold an advantage with the increased speed at which they can respond to competing businesses.
These changes in the business landscape were not felt in every corner of the society, though the next wave is building-up and it is bound to start crashing during 2020.