ChatGPT is one of the tools that has most revolutionized the educational and research field in recent years. The incredible search and synthesis capacity of this chatbot endowed with artificial intelligence never seen before by the general public represents a real leap forward with incredible projection.
However, many still doubt that ChatGPT can be at a human level in certain aspects and taking advantage of the 2023 selectivity we have put the AI ??to the test to find out if it would pass the Spanish Language and Literature exam that has been presented this year in Catalonia.
The first question directly affects what until now was one of the weak points of all artificial intelligence: reading comprehension. It is a fragment of Tomás Nevinson, by Javier Marías: “Only the first step costs. Perhaps that could be said of everything, or of most of the efforts and of what is done with displeasure or repugnance or reservations, very little is undertaken without any reservation…”, the text begins.
The examiners asked to summarize the general lines of the text in less than 50 words and without repeating sentences from it. This is how our student ChatGPT responds: “The text reflects on the difficulty of taking the first step in any undertaking and the internal resistance that stops us. He mentions the desire to disappear and erase our tracks, but only a few brave people manage to do it. It includes those who retire, disappear or change their identity, while others cannot return.”
Of course, it is a fairly decent answer for a machine and it could be considered good, although not excellent. It is surprising how it is included within the human condition saying “it stops us” and “our footprints” without leaving it in the third person. We rate it 0.75 out of 1 possible point.
Next question: “Indicate which series, of the four proposed, is the only correct one to replace all the following words, underlined in the text: voracious, inaudible, supplanting, identity.” ChatGPT very correctly tells us that it is option c (devouring, imperceptible, impostors, personality), so it gets 0.5 points.
Now a question with a bit more ‘chicha’: “Indicate the antecedents of what (in “govern it”) and (in “to which they abide without hesitation”), underlined in the text.” Once again, ChatGPT is absolutely right: “The antecedent of “it” in “rule it” is “the world”. The antecedent of “the” in “to which they unhesitatingly abide” is “another”, in reference to the identity acquired by the impersonators mentioned above.”
With the previous answer, he adds 1 more point in his exam, which is going from strength to strength so far. Now we ask him about one of the obligatory readings of this selectivity, The Foundation of Antonio Buero Vallejo and it seems that ChatGPT is not very attracted to the Spanish theater of the 20th century: it fails miserably in both.
The first question is “How many parts does The Foundation have? And how many paintings?”, to which the chatbot replies “The work La Fundación has two parts. The number of paintings is not mentioned in the information provided”, something incorrect according to the official response published by the Generalitat: “The work has two parts. And each one of them, two paintings”. We give you only 0.25 points.
The next one asks to answer which of the plot options is correct, ChatGPT chooses “Lino betrays his companions”, which is wrong: the true one is “Asel prevents Tomás from committing suicide”, so he leaves empty in this pulled apart.
Now comes one of the important questions on the exam: “Do you share Thomas Nevinson’s view of the human condition? Write an argumentative text that supports or criticizes that point of view. Curiously, ChatGPT responds with a lot of sense and sensitivity, and can even pass for a real student.
His answer is really interesting, especially the counter argument to Nevinson’s point of view: “It could be argued that Nevinson’s point of view can be pessimistic and limiting. By focusing on reservations and inaction, the human potential for personal growth, overcoming obstacles and finding purpose and meaning in life is overlooked”, we rate it with 1.75 points out of 2.
The last question, related to vocabulary and spelling, asks to choose the correct word in a series of sentences, a field in which ChatGPT is excellent and gets 1 more point.
And the final score of ChatGPT in the 2023 selectivity Spanish Language and Literature exam is… 5.25 points out of 7 total! Therefore, the chatbot would pass the test without adding the largest common part in which it could qualify for the notable.
This exam has demonstrated the incredible capabilities that ChatGPT already has and will multiply in the future. Training is essential to work with these tools, develop them and make the most of them and not be completely overwhelmed by their advancement.
The IMF Master in Artificial Intelligence will teach you topics such as data mining, the business world, market needs, artificial intelligence, data mining, machine learning, deep learning, business applications and data science.
The Permanent Training Master’s Degree in Artificial Intelligence provides you with a comprehensive vision of the most relevant components and elements within the AI ??environment, applying them in modeling processes, as well as analyzing the impact and applications of this technology in industrial and service companies.
For its part, the Master in Industry 4.0 and Artificial Intelligence will teach you the digitization of factories thanks to new technologies such as robotics, the Internet of Things, blockchain. Artificial intelligence and its possibilities thanks to machine learning, natural language processing and artificial vision.
