Advertisement

AI & SOCIETY

pp 1–11 | Cite as

15 challenges for AI: or what AI (currently) can’t do

  • Thilo HagendorffEmail author
  • Katharina Wezel
Open Forum
  • 628 Downloads

Abstract

The current “AI Summer” is marked by scientific breakthroughs and economic successes in the fields of research, development, and application of systems with artificial intelligence. But, aside from the great hopes and promises associated with artificial intelligence, there are a number of challenges, shortcomings and even limitations of the technology. For one, these challenges arise from methodological and epistemological misconceptions about the capabilities of artificial intelligence. Secondly, they result from restrictions of the social context in which the development of applications of machine learning is embedded. And third, they are a consequence of current technical limitations in the development and use of artificial intelligence. The paper intends to provide an overview of current challenges which the research and development of applications in the field of artificial intelligence and machine learning have to face, whereas all three mentioned areas are to be further explored in this paper.

Keywords

Artificial intelligence Machine learning Capabilities Data Technology development 

Notes

Acknowledgements

This research was supported by the Cluster of Excellence “Machine Learning—New Perspectives for Science” funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy—Reference number EXC 2064/1—Project ID 390727645.

References

  1. Akrich M (1995) User representations: practice, methods and sociology. In: Rip A, Misa TJ, Schot J (eds) Managing technology in society: the approach of constructive technology assessment. Pinter, London, pp 167–184Google Scholar
  2. Alcorn MA, Li Q, Gong Z, Wang C, Mai L, Ku W-S, Nguyen A (2018) Strike (with) a pose: neural networks are easily fooled by strange poses of familiar objects. arXiv:1811.11553
  3. Amodei D, Olah C, Steinhardt J, Christiano P, Schulman J, Mané D (2017) Concrete problems in AI safety. arXiv:1606.06565
  4. Ananny M, Crawford K (2017) Seeing without knowing: limitations of the transparency ideal and its application to algorithmic accountability. New Media Soc 20:973–989CrossRefGoogle Scholar
  5. Angwin J, Larson J, Mattu S, Kirchner L (2016) Machine bias: there’s software used across the country to predict future criminals. And it’s biased against blacks. https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing. Accessed 18 Jan 2018
  6. Barocas S, Selbst AD (2016) Big data’s disparate impact. Calif Law Rev 104:671–732Google Scholar
  7. Belliger A, Krieger DJ (2018) Network public governance: on privacy and the informational self. Transcript, BielefeldCrossRefGoogle Scholar
  8. Bhattacharya A (2016) Elon Musk’s OpenAI is using Reddit to teach AI to speak like humans. https://qz.com/806321/open-ai-reddit-human-conversation/. Accessed 24 Oct 2018
  9. Biggio B, Nelson B, Laskov P (2013) Poisoning attacks against support vector machines. arXiv:1206.6389
  10. Boyd D, Crawford K (2012) Critical questions for big data: provocations for a cultural, technological, and scholarly phenomenon. Inf Commun Soc 15:662–679CrossRefGoogle Scholar
  11. Brey P (2010) Values in technology and disclosive computer ethics. In: Floridi L (ed) The Cambridge handbook of information and computer ethics. Cambridge University Press, Cambridge, pp 41–58CrossRefGoogle Scholar
  12. Brundage M, Avin S, Clark J, Toner H, Eckersley P, Garfinkel B, Dafoe A, Scharre P, Zeitziff T, Filar B, Anderson H, Roff H, Allen GC, Steinhardt J, Flynn C, hÉigeartaigh S, Beard S, Belfield H, Farquhar S, Lyle C, Crootof R, Evans O, Page M, Bryson J, Yampolskiy R, Amodei D (2018) The malicious use of artificial intelligence: forecasting, prevention, and mitigation. arXiv:1802.07228
  13. Buolamwini J, Gebru T (2018) Gender shades: intersectional accuracy disparities in commercial gender classification. In: Sorelle AF, Wilson C (eds) Proceedings of machine learning research, 81st edn. PMLR, pp 1–15Google Scholar
  14. Burrell J (2016) How the machine ‘thinks’: understanding opacity in machine learning algorithms. Big Data Soc 3:1–12CrossRefGoogle Scholar
  15. Burton E, Goldsmith J, Koening S, Kuipers B, Mattei N, Walsh T (2017) Ethical considerations in artificial intelligence courses. Artif Intell Mag 38:22–36Google Scholar
  16. Campolo A, Sanfilippo M, Whittaker M, Crawford K (2017) AI now 2017 report. https://assets.ctfassets.net/8wprhhvnpfc0/1A9c3ZTCZa2KEYM64Wsc2a/8636557c5fb14f2b74b2be64c3ce0c78/_AI_Now_Institute_2017_Report_.pdf. Accessed 2 Oct 2018
  17. Carlini N, Wagner D (2018) Audio adversarial examples: targeted attacks on speech-to-text. arXiv:1801.01944
  18. Casilli AA (2017) Digital labor studies go global: toward a digital decolonial turn. Int J Commun 11:1934–3954Google Scholar
  19. Clarke R (1994) The digital persona and its application to data surveillance. Inf Soc 10:77–92CrossRefGoogle Scholar
  20. Crawford K, Joler V (2018) Anatomy of an AI system. https://anatomyof.ai/. Accessed 6 Feb 2019
  21. Cully A, Clune J, Tarapore D, Mouret J-B (2015) Robots that can adapt like animals. Nature 521:503–507CrossRefGoogle Scholar
  22. Datta A, Tschantz CM, Datta A (2015) Automated experiments on ad privacy settings: a tale of opacity, choice, and discrimination. Proc Priv Enhancing Technol 2015:92–112. arXiv:1408.6491 CrossRefGoogle Scholar
  23. Diakopoulos N (2014) Algorithmic accountability reporting: on the investigation of black boxes. Tow Center for Digital Journalism, Columbia Academic Commons, pp 1–33Google Scholar
  24. Domingos P (2012) A Few useful things to know about machine learning. Commun ACM 55:78–87CrossRefGoogle Scholar
  25. Dreyfuß H (1972) What computers can’t do: a critique of artificial reason. Harper & Row, New YorkGoogle Scholar
  26. Elgammal A, Liu B, Elhoseiny M, Mazzone M (2017) CAN: creative adversarial networks, generating “Art” by Learning about styles and deviating from style norms. arXiv:1706.07068
  27. Elish MC, Boyd D (2017) Situating methods in the magic of big data and AI. Commun Monogr 85:1–24Google Scholar
  28. Engemann C (2018) Rekursionen über Körper: machine learning-trainingsdatensätze als Arbeit am Index. In: Engemann C, Sudmann A (eds) Machine learning—Medien, Infrastrukturen und Technologien der Künstlichen Intelligenz. Transcript, Bielefeld, pp 247–268CrossRefGoogle Scholar
  29. Eykholt K, Evtimov I, Fernandes E, Li B, Rahmati A, Xiao C, Prakash A, Kohno T, Song D (2018) Robust physical-world attacks on deep learning models. arXiv:1707.08945
  30. Fang L (2019) Google hired gig economy workers to improve artificial intelligence in controversial drone-targeting project. https://theintercept.com/2019/02/04/google-ai-project-maven-figure-eight/. Accessed 13 Feb 2019
  31. Fink M (2004) Object classification from a single example utilizing class relevance metrics. In: Advances in neural information processing systems 17 (NIPS 2004), pp 1–8Google Scholar
  32. Gebru T, Morgenstern J, Vecchione B, Vaughan JW, Wallach H, Daumeé H III, Crawford K (2018) Datasheets for datasets. arXiv:1803.09010
  33. Gitelman L, Jackson V (2013) Introduction. In: Gitelman L (ed) “Raw Data” is an oxymoron. The MIT Press, Cambridge, pp 1–14Google Scholar
  34. Goldsmith J, Burton E (2017) Why teaching ethics to AI practitioners is important. ACM SIGCAS Comput Soc 110–114Google Scholar
  35. Goodfellow I, McDaniel P, Papernot N (2018) Making machine learning robust against adversarial inputs. Commun ACM 61:56–66CrossRefGoogle Scholar
  36. Goswami G, Ratha N, Agarwal A, Singh R, Vatsa M (2018) Unravelling robustness of deep learning based face recognition against adversarial attacks. arXiv:1803.00401
  37. Haggerty KD, Ericson RV (2000) The surveillant assemblage. Br J Sociol 51:605–622CrossRefGoogle Scholar
  38. Hendrycks D, Dietterich TG (2018) Benchmarking neural network robustness to common corruptions and surface variations. arXiv:1807.01697
  39. Horkheimer M (2007) Zur Kritik der instrumentellen Vernunft. Fischer Taschenbuch Verlag, FrankfurtGoogle Scholar
  40. Horkheimer M, Adorno TW (1947) Dialektik der Aufklärung: Philosophische Fragmente. Querido Verlag, AmsterdamGoogle Scholar
  41. Humphries C (2018) Digital immortality: how your life’s data means a version of you could live forever. https://www.technologyreview.com/s/612257/digital-version-after-death/. Accessed 26 Oct 2018
  42. Hyysalo S (2016) Representations of use and practice-bound imaginaries in automating the safety of the elderly. Soc Stud Sci 36:599–626CrossRefGoogle Scholar
  43. Jagielski M, Oprea A, Biggio B, Liu C, Nita-Rotaru C, Li B (2018) Manipulating machine learning: poisoning attacks and countermeasures for regression learning. arXiv:1804.00308
  44. Jia R, Liang P (2017) Adversarial examples for evaluating reading comprehension systems. arXiv:1707.07328
  45. Jiang L, Stocco A, Losey DM, Abernethy JA, Prat CS, Rao RPN (2018) BrainNet: a multi-person brain-to-brain interface for direct collaboration between brainsGoogle Scholar
  46. Kansky K, Silver D, Mély DA, Eldawy M, Lázaro-Gredilla M, Lou X, Dorfman N, Sidor S, Phoenix S, George D (2017) Schema networks: zero-shot transfer with a generative causal model of intuitive physics. arXiv:1706.04317
  47. Karas S (2004) Loving big brother. Albany Law J Sci Technol 15:607–637Google Scholar
  48. Khan B, Gawalt JR, Cook FL (2016) Science & engineering indicators. National Science Foundation, pp 1–893Google Scholar
  49. Kitchin R (2014) Thinking critically about and researching algorithms. The Programmable City Working Paper, vol 5, pp 1–29Google Scholar
  50. Kitchin R, Lauriault TP (2015) Small data in the era of big data. GeoJournal 80:463–475CrossRefGoogle Scholar
  51. Lake BM, Ullman TD, Tenenbaum JB, Gershman SJ (2016) Building machines that learn and think like people. Behav Brain Sci 40:1–58. arXiv:1604.00289 Google Scholar
  52. Lehman J, Clune J, Misevic D, Adami C, Altenberg L, Beaulieu J, Bentley PJ, Bernard S, Beslon G, Bryson DM, Chrabaszcz P, Cheney N, Cully A, Doncieux S, Dyer FC, Ellefsen KO, Feldt R, Fischer S, Forrest S, Frénoy A, Gagné C, Le Goff L, Grabowski LM, Hodjat B, Hutter F, Keller L, Knibbe C, Krcah P, Lenski RE, Lipson H, MacCurdy R, Maestre C, Miikkulainen R, Mitri S, Moriarty DE, Mouret J-B, Nguyen A, Ofria C, Parizeau M, Parsons D, Pennock RT, Punch WF, Ray TS, Schoenauer M, Shulte E, Sims K, Stanley KO, Taddei F, Tarapore D, Thibault S, Weimer W, Watson R, Yosinski J (2018) The surprising creativity of digital evolution: a collection of anecdotes from the evolutionary computation and artificial life research communities. arXiv:1803.03453
  53. Levin S (2016) A beauty contest was judged by AI and the robots didn’t like dark skin. https://www.theguardian.com/technology/2016/sep/08/artificial-intelligence-beauty-contest-doesnt-like-black-people. Accessed 10 Sept 2016
  54. Li X, Hong X, Moilanen A, Huang X, Pfister T, Zhao G, Pietikäinen M (2018) Towards reading hidden emotions: a comparative study of spontaneous micro-expression spotting and recognition methods. IEEE Trans Affect Comput 9:563–577CrossRefGoogle Scholar
  55. Litjens GJS, Barentsz JO, Karssemeijer N, Huisman HJ (2015) Clinical evaluation of a computer-aided diagnosis system for determining cancer aggressiveness in prostate MRI. Eur Radiol 25:3187–3199CrossRefGoogle Scholar
  56. Los M (2006) Looking into the future: surveillance, globalization and the totalitarian potential. In: Lyon D (ed) Theorizing surveillance: the panopticon and beyond. Willan Publishing, Cullompton, pp 69–94Google Scholar
  57. Lyon D (2003) Surveillance as social sorting: computer codes and mobile bodies. In: Lyon D (ed) Surveillance as social sorting: privacy, risk, and digital discrimination. Routledge, London, pp 13–30Google Scholar
  58. Marcus G (2018) Deep learning: a critical appraisal. arXiv:1801.00631
  59. Markow W, Braganza S, Taska B, Miller SM, Hughes D (2017) The quant crunch: how the demand for data science skills is disrupting the job market. Burning Glass Technologies, pp 1–25Google Scholar
  60. Metz C (2017) Tech giants are paying huge salaries for scarce A.I. talent. https://www.nytimes.com/2017/10/22/technology/artificial-intelligence-experts-salaries.html. Accessed 25 Feb 2019
  61. Mittelstadt BD, Allo P, Taddeo M, Wachter S, Floridi L (2016) The ethics of algorithms: mapping the debate. Big Data Soc 3:1–21CrossRefGoogle Scholar
  62. Mittelstadt B, Russell C, Wachter S (2019) Explaining explanations in AI. In: Proceedings of the conference on fairness, accountability, and transparency—FAT* ’19, pp 1–10Google Scholar
  63. Mnih V, Kavukcuoglu K, Silver D, Rusu AA, Veness J, Bellemare MG, Graves A, Riedmiller M, Fidjeland AK, Ostrovski G, Petersen S, Beattie C, Sadik A, Antonoglou I, King H, Kumaran D, Wierstra D, Legg S, Hassabis D (2015) Human-level control through deep reinforcement learning. Nature 518:529–533CrossRefGoogle Scholar
  64. Mori M (2012) The uncanny valley. Robot Autom Mag IEEE 19:98–100CrossRefGoogle Scholar
  65. O’Neil C (2016) Weapons of math destruction: how big data increases inequality and threatens democracy. Crown Publishers, New YorkzbMATHGoogle Scholar
  66. Öhman C, Floridi L (2018) An ethical framework for the digital afterlife industry. Nat Hum Behav 2:318–320CrossRefGoogle Scholar
  67. Oudshoorn N, Rommes E, Stienstra M (2016) Configuring the user as everybody: gender and design cultures in information and communication technologies. Sci Technol Hum Values 29:30–63CrossRefGoogle Scholar
  68. Pasquale F (2015) The black box society: the secret algorithms that control money and information. Harvard University Press, CambridgeCrossRefGoogle Scholar
  69. Raley R (2013) Dataveillance and counterveillance. In: Gitelman L (ed) “Raw Data” is an oxymoron. The MIT Press, Cambridge, pp 121–146Google Scholar
  70. Ribeiro MT, Singh S, Guestrin C (2016) Introduction to local interpretable model-agnostic explanations (LIME). https://www.oreilly.com/learning/introduction-to-local-interpretable-model-agnostic-explanations-lime. Accessed 26 Oct 2018
  71. Rosenfeld A, Zemel R, Tsotsos JK (2018) The elephant in the room. arXiv:1808.03305
  72. Rouvroy A (2013) The end(s) of critique: data behaviourism versus due process. In: Hildebrandt M, Vries K de (eds) Privacy, due process and the computational turn: the philosophy of law meets the philosophy of technology. Routledge, Abingdon, pp 143–168Google Scholar
  73. Rusu AA, Rabinowitz NC, Desjardins G, Soyer H, Kirkpatrick J, Kavukcuoglu K, Pascanu R, Hadsell R (2016) Progressive neural networks. arXiv:1606.04671
  74. Samek W, Binder A, Montavon G, Bach S, Müller K-R (2015) Evaluating the visualization of what a deep neural network has learned. arXiv:1509.06321
  75. Schneier B (2018) Click here to kill everybody. W. W. Norton & Company, New YorkGoogle Scholar
  76. Shillingford B, Assael Y, Hoffman MW, Paine T, Hughes C, Prabhu U, Liao H, Sak H, Rao K, Bennett L, Mulville M, Coppin B, Laurie B, Senior A, Freitas Nd (2018) Large-scale visual speech recognition. arXiv:1807.05162
  77. Shoham Y, Perrault R, Brynjolfsson E, Clark J, Manyika J, Niebles JC, Lyons T, Etchemendy J, Grosz B, Bauer Z (2018) The AI Index 2018 Annual Report, pp 1–94Google Scholar
  78. Silver D, Schrittwieser J, Simonyan K, Antonoglou I, Huang A, Guez A, Hubert T, Baker L, Lai M, Bolton A, Chen Y, Lillicrap T, Hui F, Sifre L, van den Driessche G, Graepel T, Hassabis D (2017) Mastering the game of go without human knowledge. Nature 550:354–359CrossRefGoogle Scholar
  79. Simonite T (2016) How google plans to solve artificial intelligence. https://www.technologyreview.com/s/601139/how-google-plans-to-solve-artificial-intelligence/. Accessed 18 Oct 2018
  80. Sitawarin C, Bhagoji AN, Mosenia A, Chiang M, Mittal P (2018) DARTS: deceiving autonomous cars with toxic signs. arXiv:1802.06430
  81. Steffen W, Rockström J, Richardson K, Lenton TM, Folke C, Liverman D, Summerhayes CP, Barnosky AD, Cornell SE, Crucifix M, Donges JF, Fetzer I, Lade SJ, Scheffer M, Winkelmann R, Schellnhuber HJ (2018) Trajectories of the earth system in the anthropocene. Proc Natl Acad Sci USA 115:8252–8259CrossRefGoogle Scholar
  82. Szegedy C, Zaremba W, Sutskever I, Bruna J, Erhan D, Goodfellow I, Fergus R (2013) Intriguing properties of neural networks. arXiv:1312.6199
  83. Tramér F, Zhang F, Juels A, Reiter MK, Ristenpart T (2016) Stealing machine learning models via prediction APIs. In: Proceedings of the 25th USENIX security symposium, Austin, pp 601–618Google Scholar
  84. Whittaker M, Crawford K, Dobbe R, Fried G, Kaziunas E, Mathur V, Myers West S, Richardson R, Schultz J, Schwartz O (2018) AI now report 2018. AI Now, pp 1–62Google Scholar
  85. Wiens J, Guttag J, Horvitz E (2014) A study in transfer learning: leveraging data from multiple hospitals to enhance hospital-specific predictions. J Am Med Inform Assoc 21:699–706CrossRefGoogle Scholar
  86. Wilkens A (2018) Frauenanteil an Informatik-Studienanfängern stagniert. https://www.heise.de/newsticker/meldung/Frauenanteil-an-Informatik-Studienanfaengern-stagniert-4035140.html. Accessed 24 Oct 2018
  87. Williams RM, Yampolskiy RV (2018) Optical illusions images dataset. arXiv:1810.00415
  88. Wu X, Zhang X (2016) Responses to critiques on machine learning of criminality perceptions (addendum of arXiv:1611.04135). arXiv:1–14Google Scholar
  89. Yuan L (2018) How cheap labor drives China’s A.I. ambitions. https://www.nytimes.com/2018/11/25/business/china-artificial-intelligence-labeling.html. Accessed 30 Nov 2018
  90. Zhao M, Li T, Alsheikh MA, Tian Y, Zhao H, Torralba A, Katabi D (2018) Through-wall human pose estimation using radio signals. In: Computer vision and pattern recognition (CVPR), pp 7356–7365Google Scholar

Copyright information

© Springer-Verlag London Ltd., part of Springer Nature 2019

Authors and Affiliations

  1. 1.International Center for Ethics in the Sciences and Humanities (IZEW)University of TuebingenTuebingenGermany

Personalised recommendations