The 2023 of artificial intelligence: the ethical, technological and regulatory challenges

The year 2022 has seen a “surge of creativity” in the artificial intelligence sector. The models born in recent times, in fact, are now able to produce extraordinarily convincing texts, images and even videos, upon simple ‘human’ request.

A few months have passed since the US company OpenAI launched its ‘DALL-E 2’, a deep learning model capable of producing images starting from textual instructions. And it’s a really hot trending topic the discussion about OpenAI ChatGPT, the latest large language model that interested the web with its surprising eloquence and coherence. Now let’s see what the second part of the year has in store for us.

Language models: towards the gpt-5

The last few years have seen a steady trickle of bigger and better language models than ever before. The current point of reference is the aforementioned OpenAI ChatGPT. This latest generation chatbot is a more elegant and improved version of GPT-4, the Artificial Intelligence that has started a wave of surprising ‘linguistic imitations’. And although ChatGPT has “conquered the world” of late, all eyes are on the next big thing: GPT-5. According to experts, 2023 will be the year when the next generation of large language models will kick off. And future language models could be more than just language models.

OpenAI is interested in combining different modalities, such as image or video recognition, with text. But if we take ChatGPT’s conversational capabilities and mix them with image manipulation in a single model, we’ll get something much more powerful, like being able to ask a chatbot what’s in an image or ask it directly to generate an image. If GPT-5 is to build on this technology, expect unique linguistic and imaging power packed into one “package”. By combining language and image proficiency, next-generation AI could conceivably be able to better understand both. But, of course, there is a flip side.