Artificial intelligence (AI) has experienced exponential growth and development in recent years, particularly in the field of machine learning and natural language processing (NLP). Recent and significant developments include language models based on machine learning such as OpenAI GPT-3 1 and GPT-4 2. In the process of machine learning, huge amounts of text are processed. Positive, negative, false-positive and false-negative examples are given. The possibility of a "best hit" is allowed, so that these have become extremely efficient models, which are already capable of performing various tasks, including writing texts on any topic, summarizing some text, and even programming. 

The last publicly available service (GPT-4) significantly improved its understanding of context and nuance, allowing it to generate human-like responses. Moreover, on standardized tests of general education, essay writing, and even bar exams, GPT-4 ranked among the best students of the world. This technology development has sparked an interesting discussion about how AI can affect our cognitive abilities. GPT-5 is already in preparation and it is expected that this model will be even better, more likely significantly "smarter" than any human. 

We sincerely hope that they will be smarter at least than the people who work in the agency and who wrote the report below. Because, this report seems like the work of a new analyst who just got out of college. This is released for public consumption. We humans knew much more than this as it was happening.

All Content © Croatian Film Institute, All Rights Reserved