AI Futures and Underrepresented Cultures How Will Large Language Models Shape the Cognitive Landscape of the Future?

Authors

  • Authors

  • Baraa Ali Deeb
  • Omar Almagharebi
  • Fadi A. Zaraket

DOI:

https://doi.org/10.31430/azeh1427

Keywords:

Artificial Intligence, Digital Divide, Large Language Models

Abstract

This study examines the future of artificial intelligence (AI) through the lens of “Large Language Models” (LLMs), focusing on their implications for linguistic diversity, multilingualism, and their impact on “Underrepresented Cultures”. It addresses cognitive and cultural biases affecting societies with limited digital presence. The paper investigates strategies to mitigate data disparities among languages and proposes various scenarios for the preservation of local cultures, languages, and their associated knowledge systems. Additionally, it evaluates options available to governments and policymakers in countries with underrepre-sented languages on the global AI landscape, which is currently dominated by intense competition between AI leaders in the United States and China. The paper concludes that fostering cooperation among underrepresented countries, through the collaborative development of large language models, integrated databases, and joint research initiatives, offers the most viable path forward for preserving shared cultural and linguistic heritage.

References

  1. ينظر: إدوارد سعيد، الاستشراق: المفاهيم الغربية للشرق، ترجمة محمد العناني (القاهرة: رؤية للنشر والتوزيع، 2006.)]1978[
  2. من أجل بحثٍ أكثر تفصيالًا في أنواع التحيزات وصورها، ينظر: الهيئة السعودية للبيانات والذكاء الاصطناعي، التحيز في أنظمة الذكاء الاصطناعي: تحديات وحلول (الرياض: الهيئة السعودية للبيانات والذكاء الاصطناعي، 2025)، ص 29-14، شوهد في 2025/3/17، في: https://acr.ps/1L9zQU7
  3. Lindell Bromham et al., "Global Predictors of Language Endangerment and the Future of Linguistic Diversity," Nature Ecology & Evolution, vol. 6, no. 2 (February 2022), pp. 163-173.
  4. هوميروس، الأوديسة لشاعر الخلود هوميروس، ترجمة دريني خشبة (القاهرة/ لندن: مؤسسة هنداوي، 2021)، ص.116-115
  5. Christopher A. Faraone, "Circe's Instructions to Odysseus (OD. 10.507–40) As an Early Sibylline Oracle," The Journal of Hellenic Studies, vol. 139 (November 2019), p. 50.
  6. Cole Stryker & Eda Kavlakoglu, "What is Artificial Intelligence (AI)?" IBM, 9/8/2024, accessed on 10/2/2025, at: https://acr.ps/1L9zQDu
  7. Pamela McCorduck, Machines Who Think, 2 nd ed. (Natick, MA: AK Peters, Ltd., 2004), p. 559.
  8. Amirhosein Toosi et al., "A Brief History of AI: How to Prevent Another Winter (A Critical Review)," PET Clinics, vol. 16, no. 4 (October 2021), p. 454.
  9. Allen Newell & Herbert A. Simon, "The Logic Theory Machine: A Complex Information Processing System," IRE Transactions on Information Theory, vol. 2, no. 3 (September 1956), pp. 61-79.
  10. Michael Haenlein & Andreas Kaplan, "A Brief History of Artificial Intelligence: On the Past, Present, and Future of Artificial Intelligence," California Management Review, vol. 61, no. 4 (2019), p. 8.
  11. Toosi et al., pp. 455, 456.
  12. Haenlein & Kaplan, p. 8.
  13. واجهنا خلاال البحث تحدّيًا في ترجمة المصطلحات التقنية، نظرًا إلى فقر الأدبيات العربية التي تتناول هذه المسائل. ولتفادي تشتت الجهود وتعزيز التوافق في ميدان الترجمة، اعتمدنا لترجمة المصطلحات على: الهيئة السعودية للبيانات والذكاء الاصطناعي، معجم البيانات والذكاء الاصطناعي (الرياض: الهيئة السعودية للبيانات والذكاء الاصطناعي، 2022)، شوهد في 2025/3/17، في: https://acr.ps/1L9zQCd 16 تُعتبر "الشبكات العصبية" حجر الأساس للذكاء الاصطناعي كما نعرفه اليوم.
  14. Toosi et al., p. 452.
  15. David E. Rumelhart, Geoffrey E. Hinton & Ronald J. Williams, "Learning Representations by Back-Propagating Errors." Nature, vol. 323 (1986), pp. 533-536.
  16. Tim Mucci, "The History of Artificial Intelligence," IBM, 21/10/2024, accessed on 10/2/2025, at: https://acr.ps/1L9zR7g
  17. Keith D. Foote, "A Brief History of Deep Learning," Dataversity, 4/2/2022, accessed on 12/2/2025, at: https://acr.ps/1L9zR0y
  18. Toosi et al., p. 459.
  19. Andreas Tsamados et al., "The Ethics of Algorithms: Key Problems and Solutions," AI & Society, vol. 37 (2022), p. 222.
  20. يتنافس في لعبة "غو" (Go) خصمان يتناوبان على وضع الحجارة السوداء والبيضاء عند تقاطعات لوحة شبكية. والهدف هو السيطرة على خطة اللعبة من خلاال بناء "أقاليم" داخلها؛ إذ تتحول الحجارة المحاطة بالدوائر إلى "سجناء"، ويصبح الفائز هو اللااعب الذي يملك أكبر عدد من الأراضي والسجناء. وقد جرت مقارنة انتصار برمجية "ألفا غو" على أفضل لاعبٍ في العالم آنذاك، الكوري لي سيدول (Sedol Lee)، بمباراة الشطرنج التاريخية بين الحاسوب الفائق "ديب بلو" (Blue Deep) وبطل العالم في الشطرنج غاري كاسباروف (Kasparov Garry) في سنة.1997
  21. Jean-Christophe Baillie, "Why AlphaGo Is Not AI: Google DeepMind's Artificial Intelligence AlphaGo is a big Advance but it will not Get us to Strong AI," Spectrum,17/3/2016, accessed on 18/3/2025, at: https://acr.ps/1L9zQJ6
  22. Adnan Darwiche, "Human-Level Intelligence or Animal-Like Abilities?" Communications of the ACM, vol. 61, no. 10 (September 2018), pp. 56 - 67.
  23. Ashish Vaswani et al., "Attention Is All You Need," arXiv (2023), p. 6, accessed on 13/2/2025, at: https://acr.ps/1L9zRav
  24. Will Douglas Heaven, "Why Meta's Latest Large Language Model Survived Only Three Days Online," MIT Technology Review, 18/11/2022, accessed on 17/3/2025, at: https://acr.ps/1L9zRkh المصدر: من إعداد الباحثين.
  25. Will Douglas Heaven, "OpenAI's New Language Generator GPT-3 is Shockingly Good—and Completely Mindless," MIT Technology Review, 20/7/2020, accessed on 23/3/2025, at: https://acr.ps/1L9zQSJ
  26. لطّلااع على بقية الإجابات، ينظر: جروك، شوهد في 2025/3/27، في: https://acr.ps/1L9zRkk المصدر: من إعداد الباحثين.
  27. لطّلااع على بقية الإجابات، ينظر: جروك، شوهد في 2025/3/27، في: https://acr.ps/1L9zQZ7
  28. Dwoskin, Elizabeth & Nitasha Tiku, "Tech Companies are Scaling Back Their AI Ethics Efforts just as Concerns Grow," The Washington Post, 30/3/2023, accessed on 18/3/2025, at: https://acr.ps/1L9zQCX; Mark Ryan et al., "An AI ethics 'David and Goliath': Value Conflicts between Large Tech Companies and their Employees," AI & Society, vol. 39 (March 2022), pp. 557-572.
  29. Sarah Kreps, "The Global AI Race: Will US Innovation Lead or Lag?" Brookings, 6/12/2024, accessed on 15/2/2025, at: https://acr.ps/1L9zQII
  30. Stanford HAI staff, "Global AI Power Rankings: Stanford HAI Tool Ranks 36 Countries in AI," Stanford HAI, 21/11/2024, accessed on 15/2/2025, at: https://acr.ps/1L9zQKl
  31. Nestor Maslej et al., The AI Index 2024 Annual Report (Stanford: AI Index Steering Committee, Institute for Human-Centered AI, Stanford University, April 2024), p. 14.
  32. Bedoor AlShebli et al., "China and the U.S. Produce more Impactful AI Research when Collaborating Together," Scientific Reports, vol. 14, no. 28576 (2024).
  33. Maslej et al., p. 19.
  34. بلغ الإنفاق الفدرالي على الإلكترونيات الدقيقة التي تنضوي تحتها صناعة أشباه الموصلاات 3.9 مليارات دولار في سنة 2023:. ينظر Ibid., p. 409.
  35. لا يُقصد بأشباه الموّصلّاات في هذا السياق المواد الخام شبه الموصّلة، كالسيليكون، أو حتى الترانزيستورات البسيطة، وإنما الُم نتجات العالية التقنية، أي الرقائق كالمعالجات (CPUs)، ووحدات معالجة الرسوم المتحركة (GPUs)، والآلات التي تدخل في صناعتها.
  36. Andrew Imbrie, Elsa B. Kania & Lorand Laskai, The Question of Comparative Advantage in Artificial Intelligence: Enduring Strengths and Emerging Challenges for the United States (Washington, DC: Center for Security and Emerging Technology, 2020), pp. 5-6.
  37. The White House, "Memorandum on Advancing the United States' Leadership in Artificial Intelligence; Harnessing Artificial Intelligence to Fulfill National Security Objectives; and Fostering the Safety, Security, and Trustworthiness of Artificial Intelligence," 24/10/2024, accessed on 15/3/2025, at: https://acr.ps/1L9zRas
  38. National Security Commission on Artificial Intelligence, Final Report (Washington, DC: 2021), p. 26.
  39. Matt Sheehan, "Biden's Unprecedented Semiconductor Bet," Carnegie, 27/10/2022, accessed on 18/3/2025, at: https://acr.ps/1L9zQvT
  40. Sam Winter-Levy, "With Its Latest Rule, the U.S. Tries to Govern AI's Global Spread," Carnegie, 13/1/2025, accessed on 18/3/2025, at: https://acr.ps/1L9zQF8
  41. Konstantinos Komaitis et al. "The sovereignty trap," Geotech Cues/ Atlantic council (July 2024), accessed on 18/3/2025, at: https://acr.ps/1L9zQQM
  42. Huw Roberts, "Digital Sovereignty and Artificial Intelligence: A Normative Approach", Ethics and Information Technology, vol. 26 (October 2024), p. 70.
  43. China's DeepSeek Causes Rout Among AI-linked Stocks, Nvidia Shares Tumble," Aljazeera English, 27/1/2025, accessed on 18/3/2025, at: https://acr.ps/1L9zQFb
  44. Longjun Zhou, "A Historical Overview of Artificial Intelligence in China," Science Insights, vol. 42, no. 6 (2023), pp. 969-970.
  45. Kai-Fu Lee, AI Superpowers: China, Silicon Valley, and the New World Order (Boston: Houghton Mifflin Harcourt, 2018), p. 3.
  46. Ruby Scanlon, "Beyond DeepSeek: How China's AI Ecosystem Fuels Breakthroughs," Lawfare, 14/2/2025, accessed on 18/2/2025, at: https://acr.ps/1L9zQXC
  47. Kevin Jackson, "Breaking Down Global Government Spending on AI," HPCwire, 26/8/2024, accessed on 15/2/2025, at: https://acr.ps/1L9zQC8
  48. Ngor Luong, Zachary Arnold & Ben Murphy, Understanding Chinese Government Guidance Funds (Washington, DC: Center for Security and Emerging Technology, 2021), p. 5. المصدر: من إعداد الباحثين.
  49. Dylan Butts, "Socialist AI: Chinese Regulators are Reviewing GenAI Models for 'Core Socialist Values'," CNBC, 18/7/2024, accessed on 18/2/2025, at: https://acr.ps/1L9zR55
  50. Matt Sheehan, China's AI Regulations and How They Get Made (Washington, DC: Carnegie Endowment for International Peace, 2023), p. 17.
  51. يُصطلح على تسمية ذلك "النشاط الاجتماعي السياسي للشركات" (Activity Sociopolitical.)Corporate
  52. Barbara Apaalabono Atanga, Xunyue Xue & Anna S. Mattila, "The Impact of Corporate Sociopolitical Activism (CSA) on Brand Attitude," International Journal of Hospitality Management, vol. 107 (2022), p. 103290.
  53. Maarten Buyl et al., "Large Language Models Reflect the Ideology of their Creators," arXiv (2025), p. 10, accessed on 18/2/2025, at: https://acr.ps/1L9zR9S
  54. يُعتبر التحيز في البيانات، مثال، ا من أصعب التحديات في ضبط النماذج، ومن أبرز ما يعوّق الصين عن التحكم بتدفق المعلومات؛ ما يضطر المطوّرين إلى وضع فلتر إضافي يغّير الإجابة في الوقت الفعلي. ينظر: Ryan McMorrow & Tina Hu, "China Deploys Censors to Create Socialist AI," Financial Times, 17/7/2024, accessed on 18/2/2025, at: https://acr.ps/1L9zQzx
  55. Buyl et al., p. 12.
  56. Dan Hendrycks, Eric Schmidt & Alexandr Wang, "Superintelligence Strategy," arXiv (2025), pp. 5-7.
  57. Gordon Burtch et al., "The Consequences of Generative AI for Online Knowledge Communities," Nature (May 2024), p. 10413. المصدر: من إعداد الباحثين، بالاعتماد على: LifeArchitect, "Timeline of AI and Language Models," accessed on 27/2/2025, at: https://acr.ps/1L9zQki
  58. Jared Kaplan et al., "Scaling Laws for Neural Language Models," arXiv (2020), p. 3, accessed on 20/2/2025, at: https://acr.ps/1L9zRev
  59. Tom B. Brown et al., "Language Models are Few-Shot Learners," in: NIPS '20: Proceedings of the 34 th International Conference on Neural Information Processing Systems (Red Hook, NY: Curran Associates Inc., 2020), pp. 2, 4.
  60. LifeArchitect, "Timeline of AI and Language Models," accessed on 27/2/2025, at: https://acr.ps/1L9zQki
  61. Ibid., p. 13. المصدر: من إعداد الباحثين.
  62. Alham Fikri Aji et al., "One Country, 700+ Languages: NLP Challenges for Underrepresented Languages and Dialects in Indonesia," in: Proceedings of the 60 th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) (Dublin, Ireland: Association for Computational Linguistics, 2022), p. 7226.
  63. Kaiyu Huang et al., "A Survey on Large Language Models with Multilingualism: Recent Advances and New Frontiers," arXiv (2025), p. 2, accessed on 18/3/2025, at: https://acr.ps/1L9zQFw
  64. Ibid., p. 22. المصدر: من إعداد الباحثين.
  65. Stefan Baack, "A Critical Analysis of the Largest Source for Generative AI Training Data: Common Crawl," in: Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency (New York: Association for Computing Machinery, 2024), p. 2199.
  66. تذكر الورقة نفسها التي أصدرتها "أوبن أي آي" حول نموذج "جي بي تي 3-"، أنّ اللغة الإنكليزية استحوذت على 93 في المئة من تعداد الكلمات في بيانات التدريب.
  67. Brown et al., p. 6.
  68. علمًا أن الجمع بدأ سنة 2008، ولكننا لم نستطع تمثيل كامل الإطار الزمني، بما في ذلك كامل ما جُمع خلاال سنة 2018، لأسباب تتعلق بتوافر البيانات.
  69. تقوم منهجية المكتبة البرمجية (CLD2 2, Detector Language Compact) باحتساب النسبة الاحتمالية مُرتّبة لثلااث لغاتٍ يُرجح أن تكون لغة صفحة الويب الُم.ُجرّفة
  70. Common Crawl, "Statistics of Common Crawl Monthly Archives," Common Crawl (2025), accessed on 27/2/2025, at: https://acr.ps/1L9zQzY المصدر: من إعداد الباحثين.
  71. Fangru Lin et al., "One Language, Many Gaps: Evaluating Dialect Fairness and Robustness of Large Language Models in Reasoning Tasks," arXiv (2025), p. 3, accessed on 27/2/2025, at: https://acr.ps/1L9zR3t
  72. Kate Crawford, Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence (New Haven: Yale University Press, 2021). المصدر: من إعداد الباحثين.
  73. Aleksandar Petrov et al., "Language Model Tokenizers Introduce Unfairness between Languages," arXiv (2023), pp. 10-15, accessed on 27/2/2025, at: https://acr.ps/1L9zQnW
  74. قسّم الُم (Tokenizer ستخدم في الجدول هوُ) المُ (cl100k_base) المستخدم في "جي بي تي 4-". فعلى سبيل المثال، بالنسبة إلى كلمة بسيطة مثل "لماذا"، فإن المرمز المستخدم في "جي بي تي 2-" يُقسّمها إلى خمسة أجزاء، لكلّ حرفٍ جزء: 2 p..Ibid.,
  75. Artur Kiulian et al., "From English-Centric to Effective Bilingual: LLMs with Custom Tokenizers for Underrepresented Languages," arXiv (2024), p. 1, accessed on 27/2/2025, at: https://acr.ps/1L9zQRC
  76. Muath Alduhishy, "Sovereign AI: What it is, and 6 Strategic Pillars for Achieving it," World Economic Forum, 25/4/2024, accessed on 27/2/2025, at: https://acr.ps/1L9zRdv
  77. الإمارات العربية المتحدة: معهد الابتكار التكنولوجي يشارك نموذج 'فالكون 40 بي' بشكل مفتوح المصدر لأغراض البحث والاستخدام التجاري"، معهد الابتكار التكنولوجي، 2023/5/25، شوهد في 2025/3/18، في: https://acr.ps/1L9zR7h
  78. Russia's Sberbank Releases ChatGPT Rival GigaChat," Reuters, 24/4/2023, accessed on 27/2/2025, at: https://acr.ps/1L9zQtx
  79. The Era of Sovereign LLMs and its Implications in the AI World: Insights from Cypher 2024," Machine Hack Gen AI, 29/11/2024, accessed on 27/2/2025, at: https://acr.ps/1L9zQM9
  80. Tae Kim, "Nvidia CEO Jensen Huang Predicts 'Hyper Moore's Law' Pace for AI," Barron's, 7/11/2024, accessed on 21/2/2025, at: https://acr.ps/1L9zRgy
  81. Thomas A. Campbell, "Moore's Law 2.0?" Atlantic Council, 5/9/2014, accessed on 21/2/2025, at: https://acr.ps/1L9zQV7 المصدر: من إعداد الباحثين.
  82. Gary Marcus, "CONFIRMED: LLMs have Indeed Reached a Point of Diminishing Returns," Marcus on AI, 10/11/2024, accessed on 21/2/2025, at: https://acr.ps/1L9zR48
  83. Pablo Villalobos et al., "Will We Run Out of Data? Limits of LLM Scaling Based on Human-Generated Data," Epoch AI, 6/6/2024, accessed on 21/2/2025, at: https://acr.ps/1L9zQwI
  84. Hao Chen et al., "On the Diversity of Synthetic Data and its Impact on Training Large Language Models," arXiv (2024), p. 11, accessed on 27/2/2025, at: https://acr.ps/1L9zR9E
  85. Alun Rhydderch, "Scenario Building: The 2x2 Matrix Technique," Futuribles International (June 2017), p. 16.
  86. Kevin P. Scannell, "The Crúbadán Project: Corpus Building for Under-Resourced Languages," Cahiers du Cental, vol. 5 (2007), pp. 4-5.
  87. Tianyang Zhong et al., "Opportunities and Challenges of Large Language Models for Low-Resource Languages in Humanities Research," arXiv (2024), pp. 21-22, accessed on 27/2/2025, at: https://acr.ps/1L9zRbZ
  88. His Excellency the Prime Minister Unveils Fanar, a Breakthrough in Arabic AI," Government Communication Office, 11/12/2024, accessed on 22/2/2025, at: https://acr.ps/1L9zQU9
  89. Regina Ta & Nicol Turner Lee, "How Language Gaps Constrain Generative AI Development," Brookings, 24/10/2023, accessed on 22/2/2025, at: https://acr.ps/1L9zQVb
  90. Oleksii Sharavar, "Making AI Work for all: Bridging the Language Gap in AI by Focusing on Languages that are Underrepresented in the Digital World," GSMA, 17/1/2025, accessed on 23/2/2025, at: https://acr.ps/1L9zRgo
  91. Every Language Matters: Building a More Inclusive Digital Future," UNDP, 5/10/2024, accessed on 23/2/2025, accessed on 27/2/2025, at: https://acr.ps/1L9zQQb
  92. Joanito Agili Lopo & Radius Tanone, "Constructing and Expanding Low-Resource and Underrepresented Parallel Datasets for Indonesian Local Languages," arXiv (2024), p. 10, accessed on 27/2/2025, at: https://acr.ps/1L9zRfp
  93. Zhong et al., pp. 21-22.
  94. Israel Abebe Azime et al., "ProverbEval: Exploring LLM Evaluation Challenges for Low-resource Language Understanding," arXiv (2025), p. 1, accessed on 27/2/2025, at: https://acr.ps/1L9zQmm
  95. Roberts Dar ģ is et al., "Evaluating Open-Source LLMs in Low-Resource Languages: Insights from Latvian High School Exams," in: Mika Hämaläinen et al. (ed.), Proceedings of the 4 th International Conference on Natural Language Processing for Digital Humanities (Miami: Association for Computational Linguistics, 2024), p. 289.
  96. Azime et al., p. 9.
  97. Evan Hubinger et al., "Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training," arXiv (2024), pp. 2, 4, accessed on 27/2/2025, at: https://acr.ps/1L9zQSz
  98. One in Four GCC Companies Plan to Spend More than $25m on AI in 2025," The Peninsula Qatar, 20/3/2025, accessed on 23/3/2025, at: https://acr.ps/1L9zQic
غلاف العدد 9 من دورية استشراف

Downloads

PDF (Hosted in Arabic)
HTML (Arabic)

Published

2024-07-01

Article dates

Received: March 3, 2025 Revised: March 18, 2025 Accepted: March 21, 2025

Usage statistics

38 views 60 downloads

Issue

Vol. 9 No. 009 (2024): Issue 009

Section

Studies

License

Copyright (c) 2024 Istishraf for Future Studies

Creative Commons License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.