17.01.2020 | Blog Transfer Learning & Co: Five AI trends for 2020

1. Real added value of AI is used.
After the topic of artificial intelligence has been extremely hyped in recent years, companies are looking at it more and more realistically. They notice that the fears that artificial intelligence could replace or even surpass human thinking do not correspond to reality. Instead, they are now increasingly recognizing the real added value of AI: Employees are relieved of standard and routine tasks through automation and are optimally supported in their work. At the same time, they are given more freedom for tasks that require real human intelligence.

2. Augmentation becomes reality.
Companies will use this added value in more and more places to "augment" their employees, i.e. to expand their human intelligence with artificial intelligence. In more and more areas, employees will have access to AI-based software solutions that support and assist them. These could be tools, for example, that analyze contracts and other legal documents in the shortest possible time and independently extract the relevant content. This allows employees to concentrate more on assessing risks, making decisions, and advising clients.

3. Expertise is democratized.
One effect of this augmentation is the democratization of specialized knowledge. AI-based software enables employees to access the specialized knowledge available in a company. A classic use case is customer support, for example: By making the solution to technical problems quickly accessible to new employees, they will soon be able to help customers as competently and quickly as their experienced colleagues.

4. Transfer learning is used increasingly.
A technical AI trend for 2020 is the increasing use of so-called transfer learning. In this approach, neuronal networks that have already been pre-trained for a specific purpose are used as a starting point for a different task. In this way, the trained network can use what it has already learned and the new project does not have to start from scratch, which saves a lot of time and effort. Especially when using "Big Data" models to solve "Small Data" problems (i.e. tasks where there is little or no training data) this method offers advantages.

5. AI helps with data protection.
After the DSGVO had been in effect for a year in May 2019, the authorities have now put into practice their announcement that they will impose increased penalties for violations. Companies will therefore now increasingly use AI-based software solutions to comply with the DSGVO and other data protection regulations. Such tools can, for example, determine the personal data of a specific person in the event of a request for information or find out where data relevant to the DSGVO is located throughout the company so that it can be documented and cleaned up.

Conclusion: In connection with artificial intelligence, we should seize opportunities much more than constantly putting risks in the foreground. We need to take advantage of the added value that AI can really offer us: namely to give employees access to the information they need in the shortest possible time, to provide them with insights that would escape the human eye and to relieve them of standard tasks.