Law in the Internet Society

Artificial Intelligence and Education

-- By IndraDan - 22 Nov 2023

Artificial intelligence has become one of the hottest topics of discussion. Advanced language models that replicate human speech have exploded in the last year. Anxiety over how these models will change the workplace - displacing a large number of jobs - has been a heavily emphasized narrative. Yet, the future is not predetermined. Whether these models will be net negative or positive for humanity is still undecided. They will certainly disrupt and change how a number of industries operate.

One industry that language models have the potential to revolutionize is education. Khan Academy, one of the pioneers of effective online education, has already begun to introduce AI chatbots. Despite current models lacking the full technical ability to “teach”, the technological transformation of education makes it seem inevitable that Big Tech tries. It is unlikely that the free-market will wait to identify the optimal approach for AI-based education. Instead, a trial and error approach with society is more likely. Thus, it is valuable to analyze the costs and benefits of these developments.

The Potential of AI and Education

The Case for AI

In theory, AI can represent a solution to the consistent undersupply of qualified educators in America. Language models utilize pattern analysis on extremely large sets of data in order to replicate human behavior. Furthermore, the model itself is able to finely tune its responses and develop based on the feedback that it receives from the humans who engage with them. From an optimistic perspective, an AI educator provides each student with individualized attention - teaching students iteratively at their own pace. The strength of AI as a teacher stems both from catering to the specific needs of a student and the ability of software to be deployed at scale. An AI-educator may be able to recognize certain patterns across thousands of students that would allow it to more effectively solve students' questions over time. Students could progress at their own pace and explore areas of their own personal interest - rather than being forced to stay on a single curriculum. Education could step away from standardized testing (a trend already in place) and re-focus itself on actually providing students with knowledge. Again, in theory, AI could represent a means of equalizing education across the entire world.

Unfortunately, optimistic hypotheticals must be grounded with unfortunate realities. The “convenience” of entrusting education to a seemingly all-knowing AI model is less appealing when put in context. Large language models are the newest product of surveillance capitalism. These models were developed by training on the behavioral data of human users. The battle for the collection of personal data is likely to evolve with the use of AI models. Surveillance capitalists will seek to identify the best way to make the everyday user comfortable with relying on an AI chatbot for daily tasks.

What is the Danger?

The pseudo-sentience that a language model displays may mean that these products may be more proactive about re-engaging humans to the network. The relationship between social media and anxiety, depression and a number of mental health related diseases illustrates why caution is imperative. In a world where technology has continuously expanded its influence for two decades, standing up against unwanted developments is often the hardest thing to do. Yet, Google Glass demonstrates that it is not impossible to stop the introduction of certain types or uses of technology. The discussion of AI and education requires acknowledgement of the unique role teachers play. Teachers can guide students through a variety of developmental processes, both through pre-defined curriculum and through social interactions with the other members of the classroom. This perspective on teachers indicates that they are not merely information-dispensers, but instead occupy a role in a student’s life much more akin to a “generally available mentor”. It is obvious that not all teachers are created equal and that some teachers exclusively attempt to transmit their curriculum. While not all teachers will mentor, each will hold the influence and will be positioned to advise their students.

What is a Teacher?

The reality of teaching is that most students are children - developing in every sphere of their lives and making mistakes for the first time. While the curriculum may anchor a teacher’s class time, engaging with the social dimensions of students’ lives is often impossible to avoid. Teachers (as a collective) hold immense influence over this developmental time period. Instead of focusing on the content filled in the curriculum, a successful teacher teaches their students how to think. From organization to self-reflection, students are provided with a multitude of opportunities to develop their thought processes under the tutelage of an educator. Though the connection and interaction between teacher and student has changed over time, this special human relationship has existed since ancient times.

Conclusion

It therefore begs the question - can an AI actually step into the role of a teacher? Here are a few areas of discussion that should be considered: Will an AI be able to recognize children dealing with emotional struggles, rather than more traditional academic ones? How will children be emotionally stunted by interacting (and learning) from the computer? Will AI exacerbate the teaching shortage in America? Would AI undermine the authority of human teachers in the classroom? How will the AI communicate with students?

It is interesting to think about partial solutions. Maybe certain process oriented subjects, like arithmetic, will translate better to AI education than more subjective humanities classes. Importantly, we as humans must determine whether we will allow an AI educator to step into the actual role of a teacher. Do we want an AI to teach students how to think? Or do we prefer the AI to act as an information-dispenser?

Zuboff may argue that the digital network has long been set up to modify how humans “think”. The direct access to molding human thinking would not only ensure that control, but would also develop an emotional security with technology that has never existed before. Analysis of these issues leaves more questions than answers - though, it is clear that passivity will only lead to another “dispossession cycle”.

I think the way forward is focus. Granting that "AI" exists, but only for a moment, it would be useful to get technically precise about what the "robot teacher" actually does

Whether large language models are intelligent seems relevant to whether they are "teachers," but not to whether they are "teaching." Perhaps attention to that distinction would be helpful. It will, at any rate, allow you to bypass some rather blowsy commentary on the fact that teachers and students are people.

Why do you join in the habitual extrapolation of technology you do not in detail understand? Software cannot actually drive cars in the real world. Is it more or less reasonable to suppose that software can teach, first grade or law school? What does our belief that either is possible tell us about our understanding of (a) driving, (b) teaching, and (c) software? On the other hand, I learn from reading and executing software all the time (learning by executing without reading is possible, but different learning occurs). I mostly taught myself how to do that. Does that mean that the computers I used taught me? How could they have done that if they weren't artificially intelligent? The philosophic problems you sport with here are real, and you might want to take them more seriously.


You are entitled to restrict access to your paper if you want to. But we all derive immense benefit from reading one another's work, and I hope you won't feel the need unless the subject matter is personal and its disclosure would be harmful or undesirable. To restrict access to your paper simply delete the "#" character on the next two lines:

Note: TWiki has strict formatting rules for preference declarations. Make sure you preserve the three spaces, asterisk, and extra space at the beginning of these lines. If you wish to give access to any other users simply add them to the comma separated ALLOWTOPICVIEW list.

Navigation

Webs Webs

r2 - 10 Dec 2023 - 14:24:00 - EbenMoglen
This site is powered by the TWiki collaboration platform.
All material on this collaboration platform is the property of the contributing authors.
All material marked as authored by Eben Moglen is available under the license terms CC-BY-SA version 4.
Syndicate this site RSSATOM