[Paper Review] A Glimpse in ChatGPT Capabilities and its impact for AI research

Frank Joublin, Antonello Ceravola+4 more|arXiv (Cornell University)|May 10, 2023

Artificial Intelligence in Healthcare and Education13 citations

TL;DR

The paper surveys GPT-3.5/3.4 and GPT-4 capabilities, discusses the cost and resource implications for AI research, and argues that the broad range of abilities signals progress toward general intelligence with downstream societal impact.

ABSTRACT

Large language models (LLMs) have recently become a popular topic in the field of Artificial Intelligence (AI) research, with companies such as Google, Amazon, Facebook, Amazon, Tesla, and Apple (GAFA) investing heavily in their development. These models are trained on massive amounts of data and can be used for a wide range of tasks, including language translation, text generation, and question answering. However, the computational resources required to train and run these models are substantial, and the cost of hardware and electricity can be prohibitive for research labs that do not have the funding and resources of the GAFA. In this paper, we will examine the impact of LLMs on AI research. The pace at which such models are generated as well as the range of domains covered is an indication of the trend which not only the public but also the scientific community is currently experiencing. We give some examples on how to use such models in research by focusing on GPT3.5/ChatGPT3.4 and ChatGPT4 at the current state and show that such a range of capabilities in a single system is a strong sign of approaching general intelligence. Innovations integrating such models will also expand along the maturation of such AI systems and exhibit unforeseeable applications that will have important impacts on several aspects of our societies.

Motivation & Objective

• Motivate study of large language models (LLMs) in AI research amid increasing GAFA-scale investments.
• Demonstrate the breadth of capabilities of ChatGPT-like models within a single system.
• Highlight resource and hardware costs that constrain non-GAFA research labs.
• Discuss potential innovations and unforeseeable applications as LLMs mature.

Proposed method

• Provide examples of using GPT-3.5/ChatGPT 3.4 and GPT-4 in research tasks.
• Qualitatively analyze the range of capabilities and their implications for research practice.
• Argue that the breadth of capabilities indicates approaching general intelligence.
• Discuss how evolving AI systems may spawn unforeseen applications and societal impacts.

Experimental results

Research questions

• RQ1 What capabilities do GPT-3.5/3.4 and GPT-4 demonstrate in research settings?
• RQ2 What are the resource and hardware costs associated with training and running these models for research labs?
• RQ3 How do the capabilities of these models influence AI research directions and innovation trajectories?
• RQ4 Do the observed capabilities suggest progress toward general intelligence and what are the broader societal implications?

Key findings

• The models exhibit a broad range of capabilities that can be harnessed for research tasks.
• A single system like ChatGPT shows diverse functionalities spanning translation, generation, and reasoning.
• The breadth of capabilities is interpreted as a strong sign of approaching general intelligence.
• Innovations integrating such models are expected to expand with maturation and yield unforeseeable applications.