Morocco is positioning itself in artificial intelligence [Par Charaf Louhmadi]

Artificial intelligence, machine learning, and neural networks In computer science, we call “machine learning” programs that have the ability to “learn” without modifying any line of their code. A machine learning script uses inputs or “inputs” and outputs or “outputs” to produce a method or technique to get answers that are more accurate and better... The popular Chatbot ChatGPT chat agent, developed by Californian company OpenAI, uses machine learning (machine learning) and deep learning (deep learning) to produce output. Companies are using these technologies to reduce their costs or improve and automate their internal processes, and even developers are using them to reduce costs and automate tasks. However, these tools are limited when it comes to acute academic issues, such as making complex presentations of certain mathematical theories. This can lead to bias bias in the data they are trained on and lead to discrimination in the field of recruitment and human resources. The usefulness of artificial intelligence, whose uses are increasing exponentially, has been proven by researchers and intellectuals, with the Chatbot chatbot sometimes displaying errors or taking a long time to respond to requests.

Publicerad : 2 år sedan förbi admin i Tech

Thus, machine learning programs are opposed to so-called “classical” programs based on “input” in order to produce “output”. Machine learning is mainly based on techniques and branches of applied mathematics, such as statistics, numerical methods, or optimization. The popular Chatbot ChatGPT developed by Californian company OpenAI uses both machine learning (machine learning) and deep learning (deep learning).

AI conversation agents such as ChatGPT also rely on artificial neural networks, whose creation dates back to the 1950s and whose initial goal is to mathematically model the behavior of biological neurons. In this sense, studies first targeted neurons in the mammalian cortex.

Famous psychologists, such as Frank Rosenblatt, have worked on this topic, which also bridges the gap between mathematics, biology, and psychology. The first neural network was discovered by Rosenblatt in 1957 which used very simple linear and binary models which made it understandable and practical but limited.

Also read | Enhancing the Attractiveness of Regions: A New Covenant Pattern to Stimulate Investment

More and more AI technologies and tools are using “deep learning” or deep learning, the latter involving classification of advanced and complex neural networks along with sophisticated algorithms as well. More experienced people will recognize terms like “LSTM” (long term memory), “CNN” (convolutional neural networks), or even “capsule networks”, or capsule networks.

It should be noted that techniques such as deep learning are not perfect because they encounter problems of convergence and error reduction, among others due to the size of the neural networks on board. These shortcomings can explain the nonsense that tools like ChatGPT sometimes make.

On the usefulness of artificial intelligence

Artificial intelligence is already affecting our daily lives and will continue to have a significant impact. Companies are starting to use these technologies to reduce their costs or to improve and automate their internal processes. Big tech companies like Snap or Shopify use ChatGPT APIs, for example, which are a worldwide success because they allow creating content (texts, computer programs, etc.) of more or less quality in a short time. register.

Even developers are working on it and are using extensively, sometimes on complex topics, the pieces of code that this chat agent offers. With artificial intelligence, time-consuming tasks become achievable in seconds or even minutes. We can mention, for example, the creation and design of websites.

Many employees and business people have also started getting permanent virtual assistants to delegate daily tasks to them. Students have also recognized the potential of these AI systems and used them in their presentations and work, although today these intelligences are not fully developed and cause significant errors in the answers formulated.

However, a large number of researchers and intellectuals agree to record the amazing progress of artificial intelligence and its impact on our daily lives. Thus, the usefulness of artificial intelligence, whose uses are increasing exponentially, has been proven.

To illustrate some of the existing and potentially perfect AI limitations, let’s take the example of ChatGPT. First of all, the public version of the tool uses a database discontinuing in September 2021. Second, the tool appears to be limited when it comes to acute academic issues, particularly when making complex presentations of certain mathematical theories. Thus he was unable to state from beginning to end the brilliant mathematician Paul Erdos’s proof of prime number theory.

In addition, ChatGPT does not cite sources in its answers and sometimes invents references that do not exist, which is naturally problematic with regard to its reliability. In addition, the OpenAI chatbot sometimes displays errors or takes a long time to respond when difficult requests are directed to it, and it’s not impossible to see error messages like “ChatGPT is now in capacity” or even “network error”.

Another important point, which can be seen as a limitation and threat to AI, is its sensitivity to biases in the data they are trained on. This can lead to discrimination, for example in the field of recruitment and human resources. It is therefore essential that the designers and users of these tools be aware of this and take this into account.

Also read | Qatar Airways resumes flights to Casablanca and Marrakech

Finally, AIs may not interpret certain forms of humor or sarcasm, since they are not human, or they may not have a critical mind.

One of the main threats to artificial intelligence is the destruction of jobs, sometimes even highly qualified jobs. According to recent studies, more than a quarter of jobs in Europe and the United States can be replaced fairly quickly by robots that use generative AI. Artificial intelligence has already claimed its first victims in the stock market. California-based EdTech Chegg has seen more than half of its listed value evaporate, and its president Dan Rosensweig says it’s directly related to students’ enthusiasm for ChatGPT.

Another threat from artificial intelligence: their vulnerability to cyberattacks and thus manipulation. For example, one can imagine the massive damage an autonomous vehicle, piloted by artificial intelligence and the target of a cyberattack, would cause, or if AI-powered robots in hospitals would fall victim to it. Also in this record, we can also mention the inevitable evolution of phishing emails into scams. Therefore, the strict strengthening of AI security systems is a major topic.

In the military field, if intelligent robots deployed on missions fall victim to cyberattacks, the lives of soldiers and civilians will obviously be in danger. It is not impossible to imagine in a few decades, or even a few years, complex wars in which whoever has the most detailed AI will have a huge advantage.

Moroccan companies are embarking on the use of artificial intelligence

Innovative startups, whose core business is artificial intelligence, are springing up in Morocco, and they are often launched “from zero” Thanks for the real work of the Benedictines. In health, for example, two startups, namely DeepEcho and Sobrus, aspire to revolutionize the services. These two promising companies participated in the first edition of Investing In Innovation (i3), a continental program supported by the Bill & Melinda Gates Foundation as well as the Impact Lab, a catalyst for African innovation.

The country is also increasingly interested in artificial intelligence. The Ministry of Labor will establish an artificial intelligence observatory to improve the orientation process for job seekers. Major Moroccan groups are also gradually integrating AI into their operations. Such is the case of the Moroccan multinational Intelcia, which now includes these technologies in its outsourcing operations.

Also read | An alarming rise in cash: a constant flight of bank deposits

For its part, the Mohammed VI Polytechnic University of Benguerir is participating and investing in the field of artificial intelligence. It has allocated a large budget for artificial intelligence and established an international center for artificial intelligence at the end of 2022. Last January, UM6P organized the second edition of TechInnov Day “AI for Africa” ​​and invited a world leader in artificial intelligence, Yann le Cun, award-winning Turing Award in 2018 and Director of Research and Development. . This event is jointly organized by Ecole Polytechnique and Ecole Centrale Casablanca. In light of this tangible concern of private institutions, companies and the state, the Kingdom appears to be well positioned towards these critical technological issues of the twenty-first century.y a century.

On social networks, Moroccans craze for practical websites using artificial intelligence

Several videos, which can be found on social networks, illustrate the growing enthusiasm of Moroccans for artificial intelligence. In these videos, Moroccan influencers showcase many useful websites that allow interested parties to improve their work and research. Thus, we find Moroccans praising artificial intelligence sites such as:

– www.roomgpt.io which works as an interior design tool

These clips, whose authors often express themselves in “darija”, have a certain success as evidenced by the number of views and reactions.

---

Ämnen: AI

Read at original source