ASL Lessons | Bookstore | Library | ASL University Main ►
My Brain in Exchange for Your Services: Robots as Sign Language Teachers:
Nicoletta "Nicky" Berti:
August 2024
Introduction
Artificial Intelligence (AI) evokes polarized responses, with some perceiving it as a source of fear and others as a cause for excitement. The emergence of ChatGPT has integrated AI into everyday life, generating both enthusiasm for its utility and concern over the potential development of a HAL-like entity reminiscent of 2001: A Space Odyssey. Science fiction frequently portrays robots evolving from obedient servants into malevolent rulers. Beyond fears of AI gaining consciousness and dominating the world, there are also significant concerns regarding job displacement due to AI's advanced capabilities.
Another pivotal moment in human history may be unfolding. The Industrial Revolution dramatically reshaped society, initiating an era of exponential technological advancement. Today, reality has surpassed what was once imagined. In the field of language education, the question of whether machines could replace teachers was initially dismissed due to the inaccuracies and inconsistencies in early tools like Google Translate. Despite technological improvement, teachers never seriously feared obsolescence. While it seems unlikely that machines will replace teachers, the possibility cannot be dismissed outright.
Sign language presents unique challenges due to its reliance on visual cues and spatial decoding, often overlooked in mainstream AI development. Additionally, the market for sign language technology is smaller compared to other languages, making it less attractive for largescale innovation. Furthermore, solutions designed primarily by hearing individuals, without meaningful involvement from the Deaf community, frequently result in ineffective outcomes and foster negative sentiments (Fox et al., 2023). Despite these challenges, ongoing efforts aim to bridge the communication gap between Deaf and hearing individuals, addressing the language barriers that complicate societal integration (Alsharif et al., 2023). The question remains: Can AI and robots effectively fill this gap?
Can AI replace sign language teachers?
Sign language relies on nonverbal cues that are structured and consciously integrated into communication. For example, raising the eyebrows is not merely a reaction but a deliberate part of conveying meaning. The tone is embedded in gestures and facial expressions, along with the emotion it conveys, all of which must be associated with the sign.
A robot can learn and replicate these elements, using technology to mimic human faces, voices, and sentences in multiple languages. However, the fluidity, warmth, and depth of human connection, rooted in historical and personal context, cannot be programmed into a machine. While machines can simulate emotions, they cannot genuinely experience them. Humans can detect inauthenticity; With machines, they know there is no real sentiment behind the communication.
The inherent lack of feelings in the heart of a machine is also what convinced eminent neuroscientists in the past about the low chances of seeing AI replace humans: “Unless it is possible to design a computer that has feelings…it will probably never be possible to design a computer with a mind” (Solms & Turnbull, 2002, pp. 70-71). However, considering the fast-growing progress, researchers now believe it may not be such a bizarre idea to see conscious machines (Solms, 2021). Self-aware AI may be able to reproduce that smooth and personal flavor of a language that requires multiple levels of coding to be decoded correctly: sign language.
The prospect of a sentient machine remains distant. Just as a car can be constructed due to a comprehensive understanding of physics—where the assembly of parts predictably functions according to physical laws—the creation of a sentient machine would require a similar level of understanding of the human brain. Despite significant advances in neuroscience, current knowledge is insufficient to replicate the brain artificially. The human brain still appears to be more than just the sum of its parts, making its full recreation an elusive goal.
Personality and the sense of self can be lost with steel bars crossing the orbit (yes, Phineas Gage),1 psychiatric disorders, or multiple brain conditions,2 only to make space for a different identity. A person with dementia may lose cognitive abilities but can still feel and behave like a human being, something a robot cannot naturally do even when brand new and in pristine conditions.
Day by day, the likelihood that AI will meaningfully surpass teachers in speed, knowledge, and the capacity to work incessantly increases. Although complex, sign language can be recognized and reproduced by an avatar. Cognitively, a machine can sign and communicate with a human—even function as an interpreter. However, will it be able to fulfill the critical emotional aspects that communication conveys? People might feel pity for their coffee maker,3 but the emotional pain triggered by sensing another person's feelings is far deeper. Moreover, their coffee maker cannot reciprocate feelings. There is no reciprocity.
Reciprocity
Studies show that humans seek reciprocity even in interactions with robots (Zonca & Sciutti, 2021). While robots can simulate empathy, they cannot truly feel it —they do not need to feel. Another key difference is the capacity for insight. Machines may learn from mistakes but lack the ability to develop true insights —they do not need to have insights. AI can learn from plain and rational human feedback and corrections — they do not need to be emotionally involved. However, humans do require emotional engagement and insights for meaningful learning.
Spock, the Vulcan from Star Trek, is a prime example of science fiction: If he were real, he would likely suffer from severe mental conditions and struggle to make even the simplest decisions (Damasio, 2005; Phelps et al., 2014). The idea that affects and cognition operate independently was debunked years ago by neuroscientists. There is “no clear delineation between brain regions underlying emotions and cognition…[there is a] modulatory role for emotion’s influence on cognition, and vice versa” (Phelps et al., 2014, p.265). Spock would also find it difficult to maintain relationships on the Enterprise. Patients with impaired access to their emotions also experience dysfunction in rational thought (Damasio, 2005).
This lack of genuine relationships and reciprocity would make learning from an AI challenging. Humans rely on unspoken agreements and use all of their senses to communicate. Sign language is conveyed through facial expressions, body movements, subtle variations in signs, and hand gestures. These gestures are not merely conventional; They carry a multileveled narrative. Signs evolve with the language, influenced by the individual’s unique style. When two people sign, they exchange information and express their identities, deeply rooted in their culture and experience.
Language is more than mechanical communication; It possesses a depth that should not be diminished or distorted by non-sentient robots. Language is the essence of culture; Through it, one can touch the core of a group with shared origins. Language is the way to form and consolidate meaningful relationships. Although humanity can be imitated, as demonstrated by the Turing Test,4 it should not be: Doing so would undermine the complexity and beauty of human imperfection.
Human learning is a reciprocal exchange. Winning gems or collecting streaks cannot compare to impressing mentors and fulfilling their existential purpose as teachers. While AI can perform flawlessly and simulate emotions to engage students, forming a mentorship with a robot would foster a deep sense of loneliness and dehumanization (by definition). It would also strip learning experiences of their true nature, making them less effective.
They Could, But They Should Not
Wherever technology leads, it is essential to recognize that AI can complement but not replace teachers. Humans inherently need other humans; They need to read and understand their interlocutor, something impossible with a computer. People would perceive only what the computer is programmed to show, as there is no depth beyond that. Sign language is particularly emblematic of the need to establish relationships with others expressed during communication.
History has seen countless situations where Deaf individuals were forced into mainstream schools and prohibited from using sign language, leading to isolation. Nevertheless, this pride and sense of belonging were observed as early as 1895 by a Boston Sunday Herald reporter, who noted that it was not a “camaraderie with others who have a physical condition” but a culture passed on historically (Humphries & Padden, 1988, p.2). This sense of identity, along with the unique bond formed between Deaf children when they sign to each other in a world that often makes them feel like outsiders, remains a salient aspect of Deaf culture today.
Language carries the history it embodies, which shines through in customs. In the past, Deaf individuals could only communicate in person; When others used the phone to bridge distances, they could not. So, when Deaf individuals crossed paths, it was an emotionally charged moment, making it difficult to end. It was hard to say goodbye—something that remains in today's culture (Vicars, n.d., Lesson 58). While a robot could replicate this interaction, it would lack the depth of participation and emotional connection; Even a hearing person fluent in sign language would only be able to scratch that surface after years and years of study.
Learning a language provides new insights into a different culture and perspective. It is a transformative process that broadens one's worldview and reduces the distance between people. Most are unaware of Deaf culture, and when they approach sign language, they often confront stereotypes born of ignorance (McDermid, 2017). People can be lectured about stereotypes, but it is by experiencing new realities that they can deeply evolve.
Teachers are not just instructors but mentors who serve the Deaf community by conveying enthusiasm and reducing ignorance: “Students learning about another group of people can compare and contrast themselves with that group. In doing so, students learn more about themselves and their own identity. Mentors can present a variety of historical perspectives on the social struggles of Deaf people” (Buisson & Salgo, 2012, p. 83). AI cannot offer this mentorship; How information is conveyed is not a minor detail. Learning from a computer cannot provide the same experience because it lacks the roots in centuries of historical transformation and development, an identity and pride that emerge from the soul.
The Coronavirus Disease 2019 (COVID-19) pandemic brought isolation and highlighted a profound reality: People need to feel close to others in person. Despite communicating with others through screens, many still felt lonely and depressed. Had robots been involved, this loneliness would likely have intensified, even if the robot was constantly available. Humans need other humans. They turn to pets when they lack human companionship. While AI replacing humans might seem intriguing, it would ultimately lead to loneliness and confusion. Culture would become dispersed and dehumanized, and signing would be reduced to something mechanical and technical.
Sign language teachers, on the other hand, make lessons personal and autobiographical. Some refer to the over-reliance on technology while reducing teachers to mere facilitators as learnification. Teachers are valuable not just for conveying information but for teaching on multiple levels (Guilherme, 2017). While AI might attempt to do the same, it would belong to an imaginary parallel world. Unless, of course, students were to discover tomorrow that Dr. Vicars was a robot, this argument would fail: Not only would he have passed the Turing Test, but he would also inaugurate a new era! However, the bet is that no robot will ever truly become Dr.Vicars: Human wit can be emulated, but it would not convey the same emotions. While emotions and empathy may not be critical for a computer's learning, they are fundamental to human learning and cannot be univocal. This gap will save teachers’ jobs until robots acquire awareness or humans drop it.
Conclusion
In conclusion, AI can indeed be integrated into schools to facilitate learning. The power and convenience that technology offers are promising as a complementary tool in education. However, there is a risk that society may be tempted to replace teachers with AI, similar to how automation has supplanted roles in other industries. In the past, television delivered culture and language into households worldwide, and the internet shortened distances between people. Today, AI is undoubtedly making learning language more accessible.5 However, progress often brings unintended consequences. Teachers may be forced to seek new employment, much like workers during the Industrial Revolution, and such changes may result in the loss of something vital. The tangible and intangible human qualities—authentic traits and the unique aspects of human interaction—could be diminished. Just as processed foods lose their essence and introduce new health risks, a language devoid of its human teachers may lose its cultural depth and richness. AI could potentially halt the natural evolution of the human brain, leading to cognitive atrophy as machines take over tasks once performed by humans. Maybe machines will even be able to program themselves by then. The actual risk lies not in AI overcoming humanity but in the possibility that humanity may inadvertently relinquish its essential qualities to technology. Having an atrophied brain and relying on avatars to perform tasks that are no longer familiar —or no longer understood — might be considered an acceptable trade-off, but perhaps not. AI will not maliciously need to engage in a war to defeat humans: Humans will have defeated themselves.6
1 Phineas Gage was a 19th-century railroad worker who survived a severe brain injury that dramatically changed his personality, providing key insights into the role of the frontal lobes in behavior and personality.
2 Degenerative brain conditions affecting the frontal cortex in particular, such as frontotemporal dementia, the umbrella of degenerative diseases that include Alzheimer's disease.
3 Yes, I am one of those people.
4 The Turing Test measures a machine's ability to exhibit human-like intelligence. It passes if a human evaluator cannot reliably distinguish between the machine and a human based on their responses. Alan Turing (1912-1954) was a physician and mathematician who played a fundamental role in decoding German’s encrypted messages during World War II by creating the first computer.
5 Although, at this point, why even learning? Machines could communicate on our behalf as well.
6 All this said, I want to add a note: I support the advancement of technology (I am a “geek” myself). However, I support responsible, educated progress. What I am witnessing online is frightening: People confuse reality with fiction, spend time mindlessly watching videos, and superficially rely on the internet. If algorithms generate information and masses who waste their time online are dictating those algorithms’ behavior, there will be a point where distinguishing between reality and manipulated reality will be impossible. Languages can be deprived of their deep meaning, words distorted, and the culture behind them may not have other options than crumbling. I love technology. As many other things…I hate how it is used.
References
Alsharif, B., Altaher, A. S., Altaher, A., Ilyas, M., & Alalwany, E. (2023).7 Deep Learning Technology to recognize American sign language alphabet. Sensors, 23(18), 7970. https://doi.org/10.3390/s23187970
Buisson, G. J., & Salgo, J. (2012). Mentorship: Mutual benefits for ASL students and Gifted students (part 1). American Annals of the Deaf, 157(2), 81–86. https://doi.org/10.1353/aad.2012.1612
Damasio, A. R. (2005). Descartes’ error: Emotion, reason, and the human brain. Penguin.
Fox, N., Woll, B., & Cormier, K. (2023). Best practices for sign language technology research. Universal Access in the Information Society. https://doi.org/10.1007/s10209- 023-01039-1
Guilherme, A. (2017). AI and Education: The Importance of Teacher and Student Relations. AI &Society, 34(1), 47–54. https://doi.org/10.1007/s00146-017-0693-8
Humphries, T., & Padden, C. (1988). Deaf in America: Voices From a Culture. Internet Archive. https://archive.org/details/deafinamericavoi00padd
McDermid, C. (2017). The Dialectic of Second-Language Learning: On Becoming an ASLEnglish Interpreter. Sign Language Studies, 17(4), 450-480.
Phelps, E. A., Lempert, K. M., & Sokol-Hessner, P. (2014). Emotion and decision making: Multiple Modulatory Neural Circuits. Annual Review of Neuroscience, 37(1), 263–287. https://doi.org/10.1146/annurev-neuro-071013-014119
Solms, M. (2021). Hidden spring: A journey to the source of consciousness. W.W. Norton
Solms, M. & Turnbull, O. (2002). The Brain and The Inner World: An Introduction to the Neuroscience of Subjective Experience. London: Karnac.
Vicars, W. (n.d.). Lesson 58 American Sign Language (ASL). Lifeprint. https://www.lifeprint.com/asl101/lessons/lesson58.htm
Zonca, J., & Sciutti, A. (2021). Does human-robot trust need reciprocity? arXiv (Cornell University). https://doi.org/10.48550/arxiv.2110.09359
7 Note: Alsharif and colleagues’ paper (2023) uses slightly improper language multiple times (I think involuntarily); I do not support it. However, I included it in this brief essay because it analyzes the difficulties encountered by existing machines and expresses the good intention of facilitating the integration of Deaf people through technology
Notes:
* Want to help support ASL University? It's easy: DONATE (Thanks!)
* Another way to help is to buy something from Dr. Bill's "Bookstore."
* Want even more ASL resources? Visit the "ASL Training Center!" (Subscription Extension of ASLU)
* Also check out Dr. Bill's channel: www.youtube.com/billvicars
You can learn American Sign Language (ASL) online at American Sign Language University ™
ASL resources by Lifeprint.com © Dr. William Vicars