Limitations of Artificial Intelligence - Pragmaticx

-

AI has undoubtedly brought immense changes to different fields especially, 'healthcare', with capabilities to recognize images and categorize data. They have helped to make the patient-doctor equation stronger and medical treatments faster and more affordable. Additionally, Surgeons employ AI to use smaller tools and more accurate incisions. Similarly, Visual nurse assistants may remind patients about the timing of the medication. Despite human assistance, not many of those major advancements have come. AI and humans are a balanced team in providing a diagnosis of health care. Thus, such magical wonders cannot be performed by AI itself.

 

1.    Writing computer programs: AI is not capable of writing apps. In the book 'The Mythical Man-Month', American computer architect 'Frederick Brooks' conveys that given the advances brought by AI, it does not have the human factor of understanding which makes it incapable of writing software. He explains that writing software is a process that requires a deep understanding of the real world, and the ability to transform those intricacies into rules. The secret to providing effective applications is the identification of bugs. Although AI can detect patterns to suggest a bug, it can't find bugs. In 2019 Solar Lezama and Josh Tenenbaum introduced Sketch Adapt which can write familiar parts of the software through pattern recognition.


Coding

Caption: Writing computer programs
Source: Forbes


2.      Creative content writing: AI can't just write creatively. While AI has created content, it cannot create content without guidelines. Natural Language Generation (NLG) is a software process that creates content from the data automatically. Enterprises use it to produce data reports, messaging communication, and portfolios. NLG produces thousands of documents more than humans do.

 

3.   Free will exercise: AI can make choices based on the deterministic program's rules which means, initial inputs decide the resulting actions. There are only two states of computing, do or do not. Infinite states will need to be present for AI to have free will, something which has not been done to date. AI cannot question its existence as human beings do, nor can AI explain their choices as human beings do. These questions are linked to philosophy and free will, which is not within the reach of AI.

 

4.     Self-Driving Cars: AI cannot build self-driving cars healthily and morally. Automotive firms are now about the search for self-driving cars. A report by BI Intelligence says 10 million self-driving cars will be on the road by 2020. Yet, all of them come with actual supervisors. Without this supervision, it would not be possible to have safety in the streets. Although some argue that the cause of accidents is human mistakes, and AI can bring more excellent road safety, AI can-not makes moral decisions.


Self Driving Car

Caption: Self Driving Car
Source: BBC


 

5.     Innovation and AI: AI is unable to bring innovations. AI can follow rules but like humans, it cannot create things from scratch. Human beings can invent scientific instruments, write songs, and theorems of mathematics. Unlike any product produced by AI, these innovations are genuine. AI uses past observations to learn a general pattern or model, which can be used to predict similar occurrences in the future. Like humans, AI cannot think out of the box.

 

Conclusion

This blog post is mainly focused upon limitations so despite its few limitations there are many significant advantages of AI. You can know about the advantages of AI plus other important aspects of AI here. Essentially, soon, Artificial Intelligence will indeed become ever more productive.

 


Comments

Post a Comment

Popular posts from this blog

Data Visualization Techniques

-
Image
Caption: Data Visualization Source:  searchenginejournal Exploration of data, understanding of data, and presenting the summary of the data in a beautiful way for further statistical analysis and interpretations are all the final outputs of data visualization. In the world of abundant data, visualization is essential frequently during the exploratory analysis and primary data cleaning processes in the field of machine learning, data science, and data mining. There are various techniques for data visualization which are further explored here: 1. Word Clouds Caption: Word clouds for words used in a report Source:  Nelson Norman Group Word Cloud is a visualization technique for texts done in static ways. In a word cloud, the most frequently used words are presented in larger font size and others are presented according to their frequency of appearances. It is used to summarize the contents of websites or text documents. There are typically 2 types of word cloud visualization: static and d
Read more

Artificial Intelligence Today, Applications, Abilities, Pros, Cons, Future

-
Image
Artificial Intelligence (AI) aka machine learning is a broad branch of computer science dedicated to developing smart machines that can perform tasks that usually involve human intelligence. Additionally, it is a multi-approach discipline however, the constant development in machine learning and deep learning are causing a paradigm shift in nearly every field of the tech industry possible. If you have tried reading books on AI, you might have found those textbooks define AI as a field of intelligent agents' study. Usually, the topic is often used to describe machines or computers that mimic the mental actions that associate with the human mind, such as learning and problem-solving. Now, the new learners have questions regarding how AI works. The answers are, AI works or automates, through the provided data, repetitive learning, and discovery. However, it is different from robotics automation which is driven by hardware. It does regular, high-volume, computerized tasks consistentl
Read more

Programming Languages For Artificial Intelligence

-
Image
With all the breathless hype of transitioning science fiction to a real-world where intelligence has been researched, tested and so far developed so much giving out the technological breakthroughs and the current state-of-art   of deep learning using artificial neural networks, the field of artificial intelligence in the field of the hard-working computer geeks, engineers and researchers. While all the milestones like Deep Blue defeating the world chess champion Garry Kasparov to machines learning to identify cats to self-driving cars hitting the streets and the lavish lists of all the research and development, the major foundation is the soul of algorithms being implemented through the programming languages.   Here’s the rundown of top programming languages for artificial intelligence be it machine learning or deep learning. 1.  Python Python being an open-source programming language boasts the most leverage proven pre-built libraries and community support in the field of artifici
Read more