3 Major Limitations of Artificial Intelligence (AI)

 

Artificial intelligence is shaping our future. Several large organizations are working on AI, and many companies are incorporating it into their products or services. However, at a Google event, Andrew Moore, vice president of Google Cloud, said that artificial intelligence (AI) is stupid.

Following this, many in the tech industry began sharing their views to clarify what exactly Andrew meant. He also added that:

“AI is outstanding at doing many things our brains can't do, but we can’t do that with general reasoning, including things like analogies or creative thinking.”

So, is artificial intelligence (AI) stupid?

While the question is interesting, Jeremy Goldman, founder of the Firebrand Group, agrees with Moore. He said that humans are completely superior to AI and we have just started creating AI programs. He also explained that we currently only use AI for two purposes.

1.     To simplify our current processes, which typically involve repetitive tasks in large quantities.

2.     Perform relatively easy operations compared to the human brain.

While other areas such as “creative thinking” or “out-of-the-box thinking” are still impossible to explain and difficult to work on. Artificial intelligence and machine learning can find and learn patterns, but they cannot become something new, think and make decisions like a human. Speaking of the present, there are three main limitations of artificial intelligence that are stopping tech giants from creating something big.

Limitations of Artificial Intelligence (AI)

1. Large AI data consumption

Data consumption is one of the main limitations of artificial intelligence. From the very beginning of any AI program, it requires data. It doesn't matter if the program is in the training phase or has entered the execution phase, its thirst for data will never be satisfied.

If you want to implement AI in a program, software robots must have some cognitive skills to become smarter over time. There are also robots with advanced cognitive skills that use technologies such as machine learning (ML), optical character recognition (OCR), natural language processing (NLP), and computer-aided process control (RPA) to extract meaning from the data contained in documents. After this, other roles come into play, such as task automation, which involves problem-solving or decision-making and this requires a huge amount of data.

2. Narrow AI

Artificial intelligence and its capabilities have already been explained by many reputable sources, but Briana Brownell, founder of PureStrategy.ai, a company that builds and deploys AI, explains its scope.

Currently, most AI applications are very, very narrow, she said. When it comes to artificial intelligence related to image recognition, we simply need a large number of examples so that our program can determine whether a photo is of a cat or a dog. But if you provide a photo of a rare jaguar or wolf, there is a chance that the program will not be able to identify it.

3. Emotional intelligence

Although AI is getting smarter every day, we have reached a point where processing power or speed is no longer a limitation. It's time to work on AI's emotional intelligence so it can communicate more like humans.

Natural Language Processing (NLP) must be efficient enough to understand what a person is trying to say and his/her emotions behind it. Simply put, AI needs to understand the context of the conversation.

The problem is that AI lacks emotional intelligence and therefore cannot categorize human feelings and moods into unique data or profiles. However, things will start to change in the next few years.

Artificial intelligence may seem stupid to many people right now, and it is absolutely, but the whole world is working on it, implementing it in different programs, exploring more possibilities, and it will get better every day. Perhaps we should be wary of this.