Artificial general intelligence (AGI) is still a concept, or at best in its early stages. But there is already a lot of discussion about it.
What are AGIs?
AGI and artificial intelligence (AI) are different. The latter perform specific activities such as the Alexa Assistant. But we all know that Alexa’s capabilities are limited.
AGI, on the other hand, can replace humans with robots. This allows AI to emulate human cognitive abilities. Consider a robot judge presiding over a complex case in court.
Examples of how AGI can be used in practice
Imagine a scenario where a patient with a tumor undergoes surgery. Later, it turned out that a robot had performed the operation. The results may have been successful, but the patient’s family and friends were surprised and uneasy about entrusting such a complicated task to a robot. Surgery requires improvisation and decision-making, and we trust human doctors to do that.
This concept is a terrifying and radical idea. Fear arises from a variety of ethical, social, and moral issues. One school of thought opposes his AGI because the robot can be controlled to perform undesirable and unethical actions.
AGI is still in its early stages, and despite disagreements, it will be a long time before we see it manifest. AGI has the same foundation as AI and machine learning (ML). Work is still ongoing around the world, and the main focus remains on several areas described below.
Data storage cost savings
Big data has drastically reduced the cost of data storage. Both AI and ML require large amounts of data. Big data and cloud storage have made data storage affordable and contributed to the development of AGI.
machine learning and deep learning
Scientists have made great strides in both ML and deep learning (DL) technologies. There have been great advances in neural networks, reinforcement learning, and generative models.
transfer learning
Transfer learning speeds up ML by applying existing knowledge to recognize similar objects. For example, a learning model learns to identify small birds based on features such as small wings, beaks, and eyes. Now, another learning model needs to identify different types of small birds that live in the Amazon rainforest. The latter model inherits the learning from the previous model instead of starting from scratch, which speeds up the learning.
You don’t see or experience AGI in new avatars that will change society at some point. The changes will gradually, slowly but surely appear in our daily lives.
Concerns about the development of AGI
The ChatGPT model has developed great features at an incredible speed. However, not everyone is completely convinced of his AGI potential. Various countries and experts stress the importance of guiding the development of ChatGPT within certain rules and regulations to ensure responsible progress towards AGI.
Reply from Italy
In April 2023, Italy became the first country to ban ChatGPT development due to data and payment information breaches. The government is also investigating whether ChatGPT complies with the European Union’s General Data Protection Regulation (GDPR) regulations, which protect sensitive data inside and outside the EU.
lack of transparency
Experts say there is no transparency in how ChatGPT was developed. Information about its development model, data, parameters and version release plans have not been made public.
The OpenAI brainchild continues to develop at an incredible speed, and you probably can’t imagine the power it has amassed. All without checks and balances. Some believe ChatGPT 5 will mark the arrival of AGI.
Anthony Aguirre, professor of physics and executive vice president of the Future of Life division at the University of California, Santa Cruz, said: The parameters of GPT-4 he has not disclosed by OpenAI, but there is no reason to think that this trend has stopped or slowed down.
Aguirre, who was behind the famous open letter, added that “only the institute itself knows what calculations it’s doing, but the trends are unmistakable.”
open letter
The open letter, signed by many industry insiders, reflects their fears and concerns about the chaotic development of ChatGPT.
The letter urges that all development of ChatGPT be halted until a robust framework is established to control misinformation, hallucinations and prejudices within the system. In fact, the so-called “hallucinations”, inaccurate responses, and biases that ChatGPT often exhibits are too obvious to ignore.
The open letter, which has been signed by Steve Wozniak and many other prominent figures, has already been signed by software developers, engineers, CEOs, CFOs, technologists, psychologists, PhD students, professors, doctors and public schools. 3,100 people, including teachers, have signed.
The government is also investigating whether ChatGPT complies with the European Union’s General Data Protection Regulation (GDPR) regulations, which protect sensitive data inside and outside the EU.
Abuse of AGI
It is frightening to think whether a few wealthy and powerful nations can develop and centralize AGI and use it to serve their own interests.
For example, it controls all personal and confidential data of other countries and communities and can wreak havoc.
AGI can be just a tool for biased behavior and judgment. And at worst, it leads to advanced information warfare.
Conclusion
AGI is still in the conceptual stage, but considering the lack of transparency and the perceived speed of progress in AI and ML, it may not be long before it becomes a reality.
It is imperative that countries and businesses pool their wisdom to build a robust framework with adequate checks and balances and guardrails.
The main purpose of this framework would be to protect humanity and prevent unethical intrusions into their lives.
