Abstract
AI, artificial intelligence, has recently made a big leap, especially in the field of ANI (artificial narrowed intelligence), meaning that now we are starting to have decent tools that can be useful in teaching. After the surge in importance of the distant learning techniques due to the COVID-19 pandemic in 2020, many educators have found themselves lost in dealing with an overwhelming excess of electronic information from their students, either via chat, email, documents, videos, or multimedia material. This chapter tries to delve into the difficulties of using affordable techniques for generating valid synthetic information such as rating homework or understanding if students are correctly following distant lessons. Since this is still an early subject, much more study and tests must be done to understand the full usability of automated AI tools in this (educational) context.
TopIntroduction
AI, Artificial Intelligence has recently made a big leap, especially in the field of ANI (Artificial Narrowed Intelligence), meaning that now we are starting to have decent tools that can be useful in teaching.
After the surge in importance of the distant learning techniques due to the COVID19 pandemic in 2020, many educators have found themselves lost in dealing with an overwhelming excess of electronic information from their students, either via chat, email, documents, videos, or multimedia material.
This chapter tries to delve into the difficulties of using affordable techniques for generating valid synthetic information such as rating homework or understanding if students are correctly following distant lessons. Since this is still very an early subject, much more study and tests must be done to understand the full usability of automated AI tools in this (educational) context. The content is separated into 6 main topics:
What is AI? Since AI is still not completely understood by most of the Educators, an extensive recap of what is AI and its history since 1950 has been provided,
Big Data and Neural Networks are heavily related to AI, and current AI successes are strongly connected with these new paradigms, so the basis for these technologies is also explained.
Where has it successfully being used? Many fields and examples where AI has been particularly strong and successful are listed.
Where and how has it being used in Education? Having a grasp of these concepts, we can then try to understand where AI and Big Data had already experimented in Education. To be sure the actual state of implementation is quite primordial so do not expect much concrete.
Teaching AI. Furtherly we will extend this concept in a section where we try to show that even if the actual techniques of direct AI can be challenging to implement in ordinary teachers' day life, we can plan to set the AI itself as a valuable teaching and experimental subject which is extremely interesting for students at any level.
Strong and Weak points of AI in Education. Finally, we try to understand which are the current criticism of the way AI is being used so far and show that AI can be useful as a general paradigm extending the already existent “coding” competencies that are going to be replaced or extended very soon with AI concepts.
Key Terms in this Chapter
Machine Learning (ML): Is specifically involved with the training of Neural Networ and evaluation of accuracy of predictions.
Deep Learning: Is a kind of ML where multiple layers of Neural Networks interconnected are used. It has proven in last years that the accuracy of predictions greatly improves.
Accuracy: AI and ML technology use algorithms to analyze data and make predictions based on that information. Although reports indicate that AI programs can be at least 95% accurate on a regular basis, AI programs cannot determine whether or not the data being analyzed is accurate, so usually overall accuracy is much lower but normally higher than 80%.
Neural Network: A promising new AI technique involving trainable sets of simple nodes (neurons). The network can be instructed to produce an acceptable numerical output giving some defined numerical input. Since text, images and almost everything can be represented by numbers NN can be used to define Algorithms in implicit ways, giving examples of what to do instead of explicitly programming every step.
Heuristic: A way to reach a non-perfect or optimal solution in a very complex problem. Heuristics are relatively simple shortcut or “rule of thumb” that can help in solving problem even if the problem is not perfectly solved.
Algorithm: A non ambiguous and clear specification of how to solve a problem or a class of problems, think to Cooking Recipes but with more complex logic.
Artificial Intelligence (AI): Intelligence apparently shown by machines. AI studies built devices able to perceive their environment and take action. Informally devices able to learn and to solve problems.
AIED (Artificial Intelligence in Education): Applying artificial intelligence technology to the field of education and using it in students’ learning at schools.
Big Data: Data sets that are too large and cannot sit on a single computer or storage device.