Generative Artificial Intelligence

AI in Context 

Artificial Intelligence is a large collection of technologies that generative AI (tools like Chat GPT) is only one small part of. If you think about Artificial Intelligence as a bubble that contains the necessary technological tools to create generative AI it would look like the image below. 

Artificial Intelligence is the outer ring that encompasses all technology that mimics, in various degrees, human thought. Machine learning is the next ring which includes technologies that can recognize complex patterns and create predictions or rules based on those patterns, like product recommendations. Inside machine learning is Deep Learning. Deep learning leverages multi-layer neural networks to perform more complex tasks and pattern recognition with less human intervention. 

Large Language Models, like Chat GPT, are a type of technology made possible by the multi-layer neural networks used in Deep Learning. There are other generative AI technologies housed within the Deep Learning sphere, but each includes the same basic components: A large amount of data/training set, model of neural networks, a generated output based on prompts or given task 

AI Technology By the Decade

Artificial Intelligences have been debated, discussed and speculated on well before 1950 and have often had a symbiotic relationship with movies, books and popular media with ideas and advances often crossing over into both realms. 

However, technologically the first traces of what we might now call AI begins mainly in the 1950s. 

1950

Alan Turing writes an article titled "Computer Machinery and Intelligence" detailing a test for computer intelligence called 'The Imitation Game' which later becomes know as the 'Turing Test' 

John McCarthy coins the phrase 'Artificial Intelligence' during a workshop at Dartmouth 

McCarthy develops LISP a programming language quickly adopted by researchers in the field of AI 

1960

ELIZA, the first 'chatbot', is created using Natural Language Processing to mimic human speech and logic 

Debates about the limitations and uses for neural networks begin to be explored. Frank Rosenblatt builds the Mark 1 Perceptron based off of neural network research, but Marvin Minsky and Seymour Papert argue that neural networks have server limitations and are not as effective or efficient as other models. 

1970

Debates about neural networks and ability of AI to meet the initial optimistic expectations leads to some lapses in interest and funding. However, the 1970's saw many more researchers enter the AI field. 

The Associate for the Advancement of Artificial Intelligence is created.

1980

The start of the 1980's saw a renewed interest in AI models and infrastructure, but by the end of the decade an 'AI Winter' took hold of the field. 

Expert models begin to hit the market and to be sold. Along with this, more government funding becomes available for AI research. 

Advances in neural networks, such as backpropagation, begin to be used and to advance the field more, however, advances could not meet the initial expectations for AI technology and limitations in hardware eventually contributed to a lack of investment and financial support for large scale AI research. 

1990

AI models and research continue to advance throughout the 1990's. Towards the end of the decade Deep Blue, built to win chess matches, successfully defeats world chess champion Gary Kasparov 

2000

Social media platforms begin to use AI based algorithms to inform their advertisement campaigns, but also general feeds 

Other tangential advances in cloud computing and 'big data' during this time also help advances in AI 

2010

AI begins to enter households through voice assistants like Siri and Alexa. 

Advances in hardware (like GPUs) made advances in neural networks more feasible and efficient and eventually becomes 'deep learning.' 

AI are built that can beat humans in reading and comprehension, GO and StarCraft

2020

LLMs like Chat GPT are released and more and more AI tools are being integrated into tools across the internet

IBM History of AI

Tableau History of AI

Coursera The History of AI