The brief.
For IP-2 we needed to describe five figures from the world of artificial intelligence and say how each of them thought the concept of intelligence could be identified. Then, we needed to describe the differences between programming languages and human languages, artificial intelligence and human intelligence, and machine learning and human learning. Finally, we needed to describe how our answers to these questions were different from answers that a machine might make.
-
![A portrait of Alan Turing.]()
Alan Mathison Turing
Turing was a British mathematician, code breaker, and early conceiver of the computer as a brain. He invented machines to store programs and used symbols to represent instructions for the computer to perform “any well-defined tasks” (Cowell, 2019). Turing predicted computers could imitate a person well enough to fool a human interlocutor.
Photo: The New York Times, https://www.nytimes.com/2019/06/05/obituaries/alan-turing-overlooked.html
-
![A portrait of John McCarthy]()
John McCarthy
McCarthy was an AI visionary who understood how to distill computing into a usable programming language, inventing LISP that spawned JavaScript and Python languages (Metz, 2011). He philosophized that machines would eventually be able to reason as humans do, standing counter to those who stop at machine learning when conceptualizing artificial intelligence (Metz, 2011).
Photo: Wired, https://www.wired.com/2011/10/john-mccarthy-father-of-ai-and-lisp-dies-at-84/
-
![A portrait of Herb Simon]()
Herb Simon
Simon was a social scientist who worked in organizational behaviour. He and his team created algorithms based on rules guiding human decision-making and theorized a comprehensive view of human problem-solving by employing computer simulation (Simon, 1978). He aimed to advance machine intelligence by programming machines to solve more loosely defined, complex problems (Lee, 1995).
Photo: Carnegie Mellon University, https://www.cmu.edu/50/founder-stories/story-simon.html
-
![A portrait on Marvin Minsky]()
Marvin Minsky*
Minsky was a mathematician who co-founded MIT’s CSAIL with John McCarthy. He theorized that computers could mimic the thought processes of the human mind and built the first wired neural network learning machine (“Marvin Minsky Biography”, 2022). Minsky’s machine intelligence considered the whole mind from parallel processing, to emotions, and consciousness itself (“Marvin Minsky Biography”, 2022).
Photo: The McGill Tribune, https://www.mcgilltribune.com/sci-tech/mcgill-remembers-marvin-minsky-79846/
-
![A portrait of Timnit Gebru]()
Timnit Gebru
Gebru is an electrical engineer who when working with deep learning, became aware of its social impacts and pivoted toward advising tech companies on algorithmic bias, prompting her firing from Google (Simonite, 2021). Gebru’s machine intelligence is not impartial. It is biased, requiring manual documenting, and evaluating patterns in data before use (Simonite, 2021).
Photo: Djeneba Aduayom for Wired, https://www.wired.com/story/google-timnit-gebru-ai-what-really-happened/
Human vs. machine language.
Machine languages are different from human ones in that they do not have morphologies or a natural linguistic evolution (Harris, 2018). Their syntax and semantics are designed and described in full (Harris, 2018). They do, however, share the same purpose of communication. Machine language is much more a matter of communicating through actions than words (Jones, 2020). This communication consists of actions users take within the entirety of their networked online environment, that then configure those environments based on those actions (Jones, 2020). In an algorithmic discussion, the machine hears our actions and speaks to us in adapting the context of the next conversation, informing our behavior.
Human vs. machine intelligence.
Human intelligence has not yet been concretely defined. but could be described as containing specific skills in varied contexts (Chollet, 2019). So far, machines can only generalize locally, problem-solving within set, given parameters. Humans are able to generalize much more as we adapt to handle new situations both in our experience and those of our forebears (Chollet, 2019). Human intelligence is also characterized by priors which help us achieve generalization. Machines lacking priors would not be able to transform experience into an ability to deal with future unknowns. Human intelligence is also ground in both conscious and unconscious processes, while machine intelligence is explicit and formalized (Crawford, 2021).
Human vs. machine learning.
Machine learning takes place two ways: by hard-coding prior knowledge into a machine or by adding significant quantities of training data to support machine efficiency in completing a specific skill (Chollet, 2019). Humans employ both priors and experience in their learning to afford them greater flexibility (Chollet, 2019). The quality of machine learning is predicated on the data quality used in its training (Heilweil, 2020). Both machines and humans go through training in order to meet learning objectives, and both are subject to the context and biases of their learning. If the data used in machine training is biased however, that bias is amplified over the area in which the machine operates (Crawford, 2021).
Me vs the machine.
In answering the above questions, I read wide rather than deep across multiple sources. I employed several higher order thinking processes that a machine might be unlikely to master. I read the required and recommended resources first, learning both relevant and irrelevant information rather than learning for the specific tasks. I evaluated relevant information for each question, weighting its importance. Some information that was irrelevant previously, became more relevant as my understanding grew. Machines are unlikely to be able to adapt to understanding when previously irrelevant information for a task becomes relevant (Chollet, 2019).
I have practiced writing and synthesizing over the course of many differently contextualized experiences. My intelligence also does not exist independently and reveals of me biases and characteristics that shape my voice. A machine might also depend on the human-made choices and structures (context) on which it was designed, but it is not capable of reflection and adapting to correcting its own bias or selecting better data for its decisions on its own (Crawford, 2021). I selected the data for my answers, reflected on it, revised how I employed it, and consciously made those decisions each step of the way. A machine might generate similar responses in cases where answers were more reported, but its procedural underpinnings are very different.
References.
Academy of Achievement. (2022). Marvin Minsky Ph.D.: Father of Artificial Intelligence. Academy of Achievement. https://achievement.org/achiever/marvin-minsky-ph-d/
Chollet, F. (2019, November 5). On the measure of intelligence. https://doi.org/10.48550/arXiv.1911.01547
Cowell, A. (2019, June 5). Overlooked no more: Alan Turing, condemned code breaker and computer visionary. The New York Times. https://www.nytimes.com/2019/06/05/obituaries/alan-turing-overlooked.html
Crawford, K. (2021). Atlas of AI: Power, politics, and the planetary costs of artificial intelligence. Yale University Press. https://doi.org/10.12987/9780300252392
Harris, A. (2018, November 1). Human language 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. A computer can make a decision faster. That doesn’t make it fair. Vox. https://www.vox.com/recode/2020/2/18/21121286/algorithms-bias-discrimination-facial-recognition-transparency
Jones, R. H. (2020). The rise of the Pragmatic Web: Implications for rethinking meaning and interaction. In C. Tagg & M. Evans (Eds.), Message and medium: English language practices across old and new media (pp. 17-37). De Gruyter Mouton.
Lee, J. A. N. (1995). Herbert A. Simon. IEEE Computer Society. https://history.computer.org/pioneers/simon.html
Metz, C. (2011, October 24). John McCarthy — father of AI and LISP — dies at 84. Wired. https://www.wired.com/2011/10/john-mccarthy-father-of-ai-and-lisp-dies-at-84/
Simon, H. (1978). Herbert Simon biographical. The Nobel Prize. https://www.nobelprize.org/prizes/economic-sciences/1978/simon/biographical/
Simonite, T. ( 2021, June 8). What really happened when Google ousted Timnit Gebru. Wired. https://www.wired.com/story/google-timnit-gebru-ai-what-really-happened/
*Marvin Minksy needs more thinking about as per this article in Science: https://www.science.org/content/article/what-kind-researcher-did-sex-offender-jeffrey-epstein-fund-he-told-science-he-died. This is beyond the scope of this assignment but I felt remiss in avoiding it.