Pacing Towards Technological Singularity

In a recent announcement of a new installation of Open AI’s popular Chat Bot, Chat GPT, now upgraded to GPT-4, this upgraded version now houses further privileges and functions than its predecessor. GPT-4, the 4th in the GPT series (Generative Pre-Trained Transformers), has recently sparked copious amounts of curiosity, doubt, and fear, with its enhanced capabilities. From an already polarizing Chat GPT, the public was quick to question what kind of drastic turn mankind is taking, whether it is for the better or for the worse.

“I believe that it isn’t a matter of whether AI is helpful or dangerous, but rather it is the way we employ it that ultimately decides the nature of its existence,” said a local computer science expert, Meenakshi Sundaram, who spoke on the matter.  

Compared to the previous iterations of chat bots, mainly Chat GPT and GPT 3, which have been trained on 20 billion and 175 billion parameters of information respectively, GPT-4 however, has been trained on and possesses an unreal 100 trillion parameters of information, linking it closer to a human-brain like computability and power. 

The question now stands: How does this model compare to its predecessors in terms of actual practicality and usage? To answer this question, one needs to understand that the main difference between the capabilities of GPT-4 and Chat-GPT, even GPT- 3, is that GPT-4 is able process images, whereas the older models were limited to user generated text. With this new feature, GPT-4 is able to process real-life visual information and is able answer prompts related to a particular image. For example, GPT 4 could scan an imputed image about a food label that is in a foreign language, and translate it into your spoken language. 

Experts in the tech industry have started to express their concerns over this current advancement in A.I., which is heading towards A.G.I (Artificial General Intelligence). One notable figure that recently relayed his response, post release of GPT-4, is Elon Musk. 

Musk’s team has stated, “Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable.” Showcasing their depth of seriousness on this matter, it has now become apparent that both experts and people alike should be more cautious with their responsibility towards overall development of Artificial Intelligence, as it is trained on Machine Learning; therefore, its responses can most often be unpredictable. 

As a result of this growing concern, it now stands that Elon Musk and his group of AI intelligence experts in various other industry leaders are calling for a 6 month grace period of no further development of AI past the levels of GPT-4.

But not every aspect of this advancement has to be seen as a threat, but rather a tool in which man can harness to his fullest capabilities, allowing us to not only reap its benefits but also maintain proper control of such sophisticated AI models. 

“We can employ these AI models to help us with for example, academic research, Management of the various faculties that go along with maintaining the school, Analyzing student success to failure margins in academics, and etc..,” said Kenieth Reiss, an educator from South Windsor who is looking at this matter from an educational lens.

Therefore, we should not only look for reasons to be cautious of these newly emerging technologies, but also to be accepting of our current reality and do the best we can to take advantage of the resources that such technologies provide.

In conclusion, although with the ever so increasing potential risks of AI spiraling out of control. The point is that it is here to stay. This means we should find a way to be less conservative and confining ourselves to the subject and should seek to explore it, to take a real chance at preventing technological singularity from spiraling out of control.