Artificial Intelligence

Artificial Intelligence (IP #2)

Question 1

Who were Alan Matheson Turing, John McCarthy, Herb Simon, Marvin Minsky and Timnit Gebru and what were their contributions to artificial intelligence? How did each identify intelligence?

Andrea’s Response: Alan Turing was a mathematician and scientist whose work is recognized as being foundational research of computer science and AI (Biography.com Editors, 2020). He developed a concept of a universal machine, eventually called the “Turing machine”; the precursor of the modern computer (Biography.com Editors, 2020). Alan explored the concept of intelligence through his “Turing Test” (Chollet, 2019).

ChatGPT's Response: “Alan Turing was a British mathematician, logician, and computer scientist. The Turing Test is a criterion for determining whether a machine exhibits human-like intelligence. According to this test, if a machine can have a conversation with a human in such a way that they cannot reliably distinguish between the two, then the machine can be considered intelligent” (OpenAI, 2023).

‍Andrea: John McCarthy was a computer scientist and cognitive scientist who co-founded the term “artificial intelligence” (John McCarthy, 2023). His work inspired logic programming and he produced the Lisp programming language which became the language of choice for AI application (John McCarthy, 2023).

ChatGPT: “John McCarthy was an American computer scientist and one of the pioneers of AI. He coined the term "artificial intelligence" and developed the LISP programming language, which became a prominent language for AI research. McCarthy's work focused on creating intelligent agents and formalizing  problem-solving processes” (OpenAI, 2023).

Andrea: Herb Simon was a computer scientist who researched the decision-making process within economic organizations (Herbert A. Simon, 2023). The interdisciplinary nature of his researched influenced perspectives in computer science, cognitive science, and more (Herbert A. Simon, 2023). Simon is best known for his theories of "bounded rationality" and "satisficing" (Herbert A. Simon, 2023).

ChatGPT: “Herbert A. Simon was an American economist, cognitive psychologist, and computer scientist. He made significant contributions to AI and cognitive psychology, particularly in the area of decision-making and problem-solving. Simon's research emphasized the importance of heuristics and bounded rationality in human intelligence. He believed that human intelligence was based on the ability to make satisfactory decisions under constraints” (OpenAI, 2023).

Andrea: Marvin Minsky was a mathematician and computer scientist, co-founder of the Massachusetts Institute of Technology’s AI laboratory (Marvin Minksey, 2023). He built one of the first neural-network learning machines. Minsky wanted machines to have a human capacity for commonsense and reasoning (BBC, 2016).

ChatGPT: “Marvin Minsky was an American cognitive scientist and computer scientist. He co-founded MIT's Artificial Intelligence Laboratory and made numerous  contributions to the field of AI, including work on neural networks and robotics. Minsky explored aspects of intelligence and believed that it could be achieved through the emulation of human cognitive processes” (OpenAI, 2023).

Andrea: Timnit Gebru is a computer scientist whose career focuses on algorithms bias and data mining (Timnit Gebru, 2023). She co-authored the ground-breaking paper that showed facial recognition software to discriminatory against women and  people of colour (Hao, 2020). Gebru’s work at Google is understood by peers in AI to be challenging mainstream AI practices (Hao, 2020).

ChatGPT: “Timnit Gebru is a researcher known for her work in ethics and bias in AI. She co-authored influential papers on bias in AI algorithms and co-founded the organization Black in AI to promote diversity and inclusion in the field. While Gebru's work is more focused on addressing the ethical and social implications of AI, her research highlights issues of fairness, accountability, and transparency in AI systems” (OpenAI, 2023).

Question 2

How do “machine (programming) languages” differ from human (natural) ones?

Andrea’s Response: Human (natural) languages are very flexible as they change and evolve depending on  many variables; quite differently, programming languages are developed with a strict set of rules (Harris, 2018). The meaning and pronunciation of words in natural language can change depending on the context (Harris, 2018). They also have: logic and emotion, and are influenced by historical and cultural references—much of which is what makes natural language beautiful and complex (Harris, 2018). Programming languages are also complex, but since their parameters are designed beforehand, are more fixed in meaning and built for performance; remaining precise and clear (Harris, 2018).

ChatGPT's Response: “Machine programming languages and human natural languages differ significantly in their structure, purpose, and usage. There are key distinctions between the two which occurs in: purpose, syntax, vocabulary, ambiguity, communication medium, and learning and use. Machine programming languages are engineered for the precise, unambiguous communication of instructions to computers, while human natural languages are evolved systems for human communication that are inherently more flexible, expressive, and context-dependent. Each serves its unique purpose and has distinct characteristics tailored to its intended use” (OpenAI, 2023).

Question 3

How does “machine (artificial) intelligence” differ from the human version?

Andrea’s Response: Machine (artificial) intelligence and human intelligence are often compared based on skill, but that comparison doesn’t effectively measure intelligence as it heavily varies based on experiences and prior knowledge (Chollet, 2019). Artificial intelligence is programmed to be powerful and its capability grows when more data is provided. Where it falls short is solving complex problems—it is not as agile as human intelligence. Unlike AI, human intelligence has the ability to evolve by turning experience into knowledge and skill that spans an endless variety of uses and applications (Chollet, 2019). Human intelligence can be highly adaptive, includes emotion, and is able to establish morals and ethics; all to uniquely develop human intelligence that can run as wide as it can deep (Human Intelligence, 2023).

ChatGPT's Response: “Machine (artificial) intelligence (AI) differs from human intelligence in several fundamental ways, reflecting both the capabilities and limitations of current AI systems; by origin, learning and adaptation, speed and accuracy, emotion and empathy, creativity and innovation, self-awareness and consciousness, and ethical and moral judgement. In summary, while machine intelligence (AI) has made remarkable progress in performing specific tasks and automating various functions, it lacks the depth, breadth, and nuanced capabilities of human intelligence. Human intelligence encompasses a wide range of cognitive, emotional, and ethical dimensions that make it a uniquely complex and adaptable phenomenon. Current AI systems are powerful tools, but they do not replicate the full scope of human intelligence and consciousness” (OpenAI, 2023).

Question 4

How  does “machine learning” differ from human learning?

Andrea’s Response: Machine learning lacks the ability to discern biased data and information that is provided to them, which, in turn, could lead to harming people (Artificial Intelligence, 2023). In addition, AI may also contain bias from how and who  develops the system—known as algorithmic bias (Heilweil, 2020). Human learning may not have as huge of a library of knowledge as machines are capable of, but human learning does have the crucial ability to rationalize though complex cognitive processes that are necessary for disseminating false and biased information.

ChatGPT's Response: “Machine learning and human learning differ in several key ways, reflecting the fundamental differences between artificial intelligence and human cognition. Differences exist in learning process, learning capacity, creativity and innovation, emotion and motivation, transfer of knowledge, feedback and self-reflection, innate knowledge, and ethical and moral considerations—AI requires human intervention for both. In summary, while machine learning is a powerful tool  for specific tasks and automation, it is fundamentally different from human learning, which is a multifaceted cognitive process encompassing emotions, creativity, abstraction, and a wide range of adaptable skills and knowledge. Human learning is deeply intertwined with consciousness and self-awareness, qualities that current machine learning systems do not possess” (OpenAI,  2023). 

Question 5

How do my answers to these questions differ from what a machine (e.g., such as  those from ChatGPT) could generate?

A key difference that stands out between my responses and ChatGPT’s are that mine contain reasoning and show understanding. Before this exercise, I really didn’t know much about AI and spent a lot of time reading and thinking to formulate an understanding. I think reasoning to form understanding also allows for emotion and creativity to enter my thought process to assist in formulating a response that is my own. I’m still making sense of everything, so my writing included lots of citations.

Involving emotion, creativity and drawing upon some life experience as I have, is not something that machine learning does. This was my first time using ChatGPT, and I was shocked by the length of response it provided, but those responses are clinical; void of deep expression. A human is capable of providing a more multifaceted  response. Looking back at my response to question number two, I noted some of the key properties of natural language and added that those elements are what make natural language beautiful and complex. I drew upon my own understanding of human language from a perspective of experience (from my decades past) and from studies about English and various other humanities. AI cannot draw upon life experience or formulate understanding, full of nuances and character, in a way that humans can.  

References 

Artificial Intelligence. (2023, September 23). In Wikipedia. https://en.wikipedia.org/w/index.php?title=Artificial_intelligence&oldid=1176759254

BBC. (2016, January 26). AI Pioneer Marvin Minsky Dies Aged 88. BBC News. https://www.bbc.com/news/technology-35409119

Biography.com Editors. (2020, July 22). Alan Turing Biography. The Biography.com. https://www.biography.com/scientists/alan-turing

Chollet, F. (2019). On the Measure of Intelligence. https://doi.org/10.48550/arxiv.1911.01547

Harris, A. (2018, October 31). Human Languages vs. Programming Languages. Medium. https://medium.com/@anaharris/human-languages-vs-programming-languages-c89410f13252

Heilweil, R. (2020, February 18). Why Algorithms Can Be Racist and Sexist. Vox. https://www.vox.com/recode/2020/2/18/21121286/algorithms-bias-discrimination-facial-recognition-transparency

Herbert A. Simon. (2023, September 20). In Wikipedia. https://en.wikipedia.org/w/index.php?title=Herbert_A._Simon&oldid=1176216604

Hao, K. (2020, December 7). We Read the Paper That Forced Timnit Gebru out of Google. Here’s What It Says. MIT Technology Review. https://www.technologyreview.com/2020/12/04/1013294/google-ai-ethics-research-paper-forced-out-timnit-gebru 

Human Intelligence. (2023, September 11). In Wikipedia. https://en.wikipedia.org/w/index.php?title=Human_intelligence&oldid=1174885158

John McCarthy (computer scientist). (2023, August 23). In Wikipedia. https://en.wikipedia.org/w/index.php?title=John_McCarthy_(computer_scientist)&oldid=1169369534  

Marvin Minsky. (2023, September 5). In Wikipedia. https://en.wikipedia.org/w/index.php?title=Marvin_Minsky&oldid=1174018575

Timnit Gebru. (2023, August 12). In Wikipedia. https://en.wikipedia.org/w/index.php?title=Timnit_Gebru&oldid=1169927648 

OpenAI. (2023). ChatGPT (GPT-3.5 version) [Large language model]. https://chat.openai.com

‍