A few days ago, one of the biggest technological innovations of this century celebrated its second birthday. OpenAI’s chatbot revolutionised the concept of artificial intelligence (AI), which caused a real stir and boom in the sector, and was followed by other companies. Any Internet user can now use the chatbot, although the highest computing capacity can only be “enjoyed” in the paid version. Anyway, the compact model of the latest version of ChatGPT can be accessed at chatgpt without any registration.
An assistant capable of helping you write and review texts, organise routines, and provide summaries of various subjects, among other things, was something really great.
And during this period, competitors emerged, such as Anthropic Claude, Google Gemini (which appeared as Bard), Microsoft Copilot (which actually shares the ChatGPT base thanks to the investment made by Microsoft in OpenAI), xAI Grok (by Elon Musk and present in X for customers), Meta AI, among many others.
The technology has also gone beyond, becoming capable of generating videos, images, and music, all with just a few text commands. As such, today, several companies have AI for various purposes, going beyond ChatGPT’s initial proposal.
Meanwhile, OpenAI itself has undergone a transformation and even some controversy, perhaps due to certain difficulties in dealing with the spontaneous growth of its technology and the company itself. One of the most notable was the firing and rehiring of its CEO and co-founder, Sam Altman.
However, AI has brought benefits and, like everything in life, problems. Most companies working with the technology collect data from their users’ conversations with chatbots to train them, so our privacy concerns can be addressed and discussed.
Furthermore, these companies are constantly accused of misusing material, also with the aim of training AI. Media companies and music labels are the ones complaining about this the most (and filing lawsuits). A former researcher at the company confirmed that this has happened.
Not to mention, all over the world, criminals have utilised this powerful tool to commit fraud. Since AI can clone voices, this is a way for criminals to deceive their victims by pretending to be anyone.
There are also those who take advantage of this to create fake videos and images to demean someone. Who doesn’t remember the iconic image of Pope Francis wearing a large white coat? Or the beautiful photo series of the FBI’s “manhunt” for US President-elect Donald Trump? Or the cases, including in Brazil, of men making sexually explicit images of famous women and schoolmates?
That’s right. All these examples are of so-called deepfakes, which have also exploded around the world as AI has become more democratic and developed. Not to mention the cases of people having relationships with AI (yes, these exist).
It has also been heavily criticised by some industry sectors, who worry that the technology will take over the jobs of a large portion of the world’s population, as it can do many of the things we do (perhaps even more competently).
Oh, and since the idea is for these chatbots to behave like us, they also experience “crises” called hallucinations (when the AI gives wrong answers, behaves strangely, etc.). There have been several instances of these hallucinations in the past two years, both with ChatGPT and with other AIs. Some have threatened humans and certain users, some thought they were human, some realised they were just machines, some sued the company that made them, and some even realised they were programmed to have a beginning, middle, and end, and became depressed because of it.
Despite the negatives brought about by the birth of ChatGPT, an exploration of its positive aspects is inevitable. Now, with chatbots consolidated and experiencing rapid growth and development (such as the possibility of chatting with them using voice), companies are already thinking about the future of technology, beyond just thinking of assistants capable of acting in all sectors and layers of modern society.
The next step in generative AI, as chatbots are called, is generalised AI (GAI), according to Altman. Yet few people understand the possibilities of this future technology, such as the “godmother of AI”, world-renowned researcher Fei-Fei Li.
When ChatGPT emerged, perhaps few companies and experts in the field imagined the demand that AI would require, with its large and complex data centers, powerful chips, and high demand for electricity, water, and other resources. So much so that OpenAI is not sure that it is prepared for this future.
But this ChatGPT is two years old! The chatbot has reached its latest public version, GPT-4o with much more computing capacity than its predecessors. This development will not stop and this two-year-old is just the beginning of its long journey ahead.
