Can Gemini AI Independently Learn Language?

Our surroundings are changing quickly in the age of artificial intelligence (AI). Doing what was unthinkable just a few years ago is now possible. The unthinkably quick development of technology and artificial intelligence has, however, also brought up some unexpected developments. Does artificial intelligence have the ability to learn anything by itself? Language is also a difficult issue. The assertion made last year that Google’s AI model Gemini (previously Bird) had picked up Bengali on its own sparked a contentious discussion among researchers and specialists in the field. However, let’s investigate whether it is plausible that an AI could pick up a language on its own or whether there is another reason.

Google’s Claim vs. Gemini’s Proficiency in Bangla

In an interview with CBS News ’60 Minutes program, Google CEO Sundar Pichai and the company’s technology leader James Manyika asserted that their AI model, Gemini, could comprehend and interpret Bangla with only a few cues. Sundar Pichai, the CEO, described this surprising occurrence as an emergent trait of AI.

Google’s assertion that Gemini picked up Bangla on its own, however, has caused a great deal of criticism. However, AI research has not yet provided a complete explanation for artificial intelligence’s capacity for self-learning.

Unraveling the mystery: What scientists revealed

AI experts swiftly disputed Google’s assertion. According to former Google researcher Margaret Mitchell, Google’s earlier AI model, Palm (PaLM), also known as Bird or Gemini, was trained on a sizable Bangla data set before its creation. In other words, Bird learned Bangla primarily from previously trained data rather than through innate learning.

According to AI specialists, what seems to be self-learning is just a pattern that Bird or Gemini, as we call it, has taken from its earlier machine-learning models. Extensive multilingual datasets are used to train huge language models like Gemini or ChatGPT. When an AI is presented with a Bengali prompt, it uses the knowledge it has already acquired to generate replies; it does not acquire a new language on its own.

AI’s “black box”

In the field of artificial intelligence, the “black box,” or unknown, is one of the most perplexing concepts. Even those who create AI are occasionally unable to comprehend how and why their models arrive at a particular result. Even Sundar Pichai, the CEO, acknowledged that AI might occasionally yield conclusions that are quite enigmatic.

Is self-learning possible for an AI?

Although AI systems frequently display unanticipated behavior, experts claim that they are not equally adept at learning as the human brain. AI just uses previous data and statistical patterns to acquire language, whereas humans learn language by experience, social contact, and logical thought. It’s not actually “self-learning” as it may appear.

The Future of Artificial Intelligence: What’s Next?

A new technological era appears to be upon us, as seen by the quick development of AI technologies from Gemini, ChatGPT, and Microsoft. Although these systems are quite good at processing behavior that is similar to that of humans, they are still not quite as intelligent as humans. But as AI develops, it’s critical to maintain accountability and openness. Technology businesses should avoid misusing or instilling fear of AI by accurately disclosing the genuine capabilities and limitations of their models. Although artificial intelligence has advanced to the point that it can learn on its own, analysts of AI think that it is unlikely to ever fully replace human intelligence and thinking.

“References”