The author of the above article, Vinod Khosla, sets out a case for algorithms ‘diagnosing diseases and teaching high school’ in the future, with less able teachers ‘providing the human touch as mentors and coaches’.
I think there is some mileage in this argument and some areas where it falls down. It’s certainly true that algorithms are already supporting the diagnosis of disease, such as those used in cervical cytology and skin lesions, and one can imagine how the accuracy of disease identification could improve as more and more varied presentations of diseased and normal tissue are added to the networked algorithms data sets. Human doctors must rely on a similar process of recalling previous presentations that have enough similarities to suggest a likely diagnosis, or consulting their more experienced colleagues and online resources, all of which is essentially algorithmic behaviour.
One could imagine how similar algorithms might help someone learn and recall facts, or learn a language, in essence anything that is binary, where there’s a right and a wrong answer. We’re already seeing such tools such as Memrise, a language learning app, being developed and made available on mobile devices.
Where I think the argument for algorithm driven teaching could fall down is, as Siemens (2013) states, because “The learning process is creative, requiring the generation of new ideas, approaches, and concepts.” Whereas “analytics, in contrast, is about identifying and revealing what already exists”. Analytics provides the data that algorithms use to solve problems, so one can see how an algorithm would struggle to understand creativity, new ideas and approaches, other than by classifying them as being like something identified correctly previously.
Earlier this week I was listening to a university professor explaining how self driving cars will become safer as they will pass on data about any collisions they do have to all the other self-driving cars. This is probably true, but an ability to identify similar sets of circumstances could never make them completely safe without removing human creativity and unpredictability from the equation by having no human drivers and no pedestrians in the vicinity of the vehicles. Similarly computerised teachers could learn and share successful teaching methods with other computers on the network, but I’m not sure how they would teach a student to make new and unexpected connections. However, technology continues to advance and this video would suggest that artificial intelligence could be given the capacity to reason, rather than complete a predetermined set of steps.
Another article I found online this week suggest a compromise might be the more immediate outcome. The article, titled “Could online tutors and artificial intelligence be the future of teaching?” suggests that a new software platform “will become one of the first examples of artificial intelligence (AI) software being used to monitor, and ideally improve, teaching.”. Tom Hooper, chief executive officer of the company who created the platform said: “We’re looking to optimise lessons based on the knowledge we gain. We’ve recorded every lesson that we’ve ever done. By using the data, we’ve been trying to introduce AI to augment the teaching”.
“Initially, the company’s 300 tutors will receive real-time, automated interventions from the teaching software when it detects that a lesson may be veering off-course.” It will be interesting to see how these tutors respond to the software effectively monitoring and feeding back on their performance in real time. I also wonder whether there will be cultural differences in how teachers in other countries will respond to the same type of input. The current tutors are based in India and Sri Lanka. The following table (Weil, M Rosen, L. 1999) either indicates that attitudes to technology in the region have progressed at pace in the past seventeen years, or the human aspect of this venture may be the more problematic.
Khosla, V. (2012) Will We Need Teachers Or Algorithms? Posted on techcrunch.com Jan 15, 2012
Siemens, G. (2013) Learning Analytics: the emergence of a discipline. American Behavioral Scientist, 57(10): 1380-1400
Devlin, H. (2016) Could online tutors and artificial intelligence be the future of teaching? https://www.theguardian.com
Weil, M Rosen, L. (1999) The psychological impact of technology from a global perspective: A study of technological sophistication and technophobia in university students from twenty-three countries, Computers in Human Behavior Volume 11, Issue 1, Spring 1995, Pages 95–133