AI

Artificial Intelligence and the Question of Consciousness

An Artificial Lawyer

This February, an artificial intelligence (AI) legal assistant is scheduled to provide guidance to a human arguing against a speeding ticket in court.[1] These services will be provided through the application DoNotPay, a program created by Joshua Browder with the goal of providing legal representation for the average citizen.[2] The AI will listen to the court proceedings through the DoNotPay app and then prompt the defendant on what to say through an earpiece. This is expected to be the first case defended by a “robot lawyer”.[3]

The simulation of human intelligence is referred to as artificial intelligence, a term first coined by John McCarthy in 1956 at Dartmouth Conference.[4] Exact definitions have varied over time, but IBM defines artificial intelligence as, “a field, which combines computer science and robust datasets, to enable problem-solving. It also encompasses sub-fields of machine learning and deep learning…”[5]

Recent years have seen an explosion in AI with versions now prominent in social media, health care, finance, ecommerce, agriculture, education, and e-marketing.[6] In 2017, Sophia, a social humanoid robot, was found to be so realistic that she was granted Saudi-Arabian citizenship.[7] Last year 2022 was considered to be a pivotal period for Generative AI, a subcategory for artificial intelligence that creates original imaginative content including artwork and stories through machine learning algorithms.[8] These rapid transitions may leave many wondering: what is coming next?

This image was created by NightCafe, an AI artist that generates images based on a description it is given. This piece was generated by the request for, “a picture of artificial intelligence.”[9]

 

Artificial Intelligence and Consciousness

Social robots (defined as “a physically embodied autonomous agent that communicates and interacts with humans on a social level”) have qualities that can make them appear human to those who interact with them including recognizing speech, communicating verbally, exhibiting behavior, learning, decision making, and facial expressions.[10] This raises the “hard problem of consciousness” referring to the difficulty of understanding whether these beings can ever be considered conscious and whether subjectivity can emerge from matter.[11]

Artificial consciousness (AC) is the aspect of the mind that is generally thought to be the most difficult to replicate or produce artificially.[12] To determine whether a machine is conscious, it is first necessary to answer the proverbial question: What is consciousness? This topic has been widely debated by philosophers for centuries and as of now no clear definition is agreed upon. In a recent New York Times’ article, Oliver Whang (2023) compares defining consciousness with other large vague concepts such as love, freedom, and existence.[13] While early philosophers Descartes and Leibniz believed that consciousness could not be explained by simple mechanics, some modern programmers disagree.[14,15]

Meissner, (2020) cites the Computational Theory of the Mind (CTM) which claims that the human mind is essentially a computer and defines consciousness as, “the awareness of one’s own existence.”[16] However, he distinguishes between “primitive consciousness,” which only has a basic understanding of one’s existence and “reflective consciousness,” which includes the ability to analyze the existence of both oneself and others.[17] This notion is reinforced by Banjeree (2020) who wrote a model for artificial consciousness based on the belief that consciousness is a system of information processing that requires a feedback loop in which a deep learning network feeds back into itself and the emergent property that is produced.[18] He argues that not only can a computer be taught to recognize itself, but it can also learn to identify its own unique characteristics that make it separate from other computers such as memories of its own past actions, color, size, identification markers, and amount of memory.[19] Xie, (2021) states that the generation of consciousness can be tracked by the development of memory, rationality, self-awareness, intelligence, sensation, and emotion.[20] This can be combined with a seventh ability of super-intelligence, which might be possessed by a potential super-human in the future.[21] However, Zhou, (2021) asserts that human consciousness requires emotions, that the motivation of rational thinking is derived from emotions, and that intelligence without emotion can neither exist or be meaningful.[22]

To further explore this question, a series of tests have been proposed to evaluate whether an artificial entity has developed consciousness. One problem encountered with this approach is that there is no corresponding definition of consciousness for other individual living entities, such as amphibians.[23] Despite the evasiveness of this type of categorization, attempts at this assessment have already begun. The wise-man puzzle (a.k.a. knowledge game or muddy children puzzle) was a test developed by Luciano Floridi in 2005 intended to challenge AI with philosophically rich and challenging questions to prove consciousness.[24] However, research conducted by Bringsjord, (2010) indicated there may be ways to program AI to pass this test.[25] Pipitone & Chella, (2021) similarly determined that a robot is capable of passing the classic mirror test (a test evaluating if a robot is self-aware enough to see itself in the mirror) by talking to itself.[26] Researchers are still working to decide if any test of consciousness exists that could not be conquered solely through programming.

Dr. Hod Lipson, who has been working on AI development since the early 2000s recently explained to the New York Times that developers take cues from nature and the process of evolution.[27] He states that whether AI can develop consciousness is not just another research question, but it is “the question” explaining that it could eclipse everything else that scientists have done.[28] A recent article published in Forbes magazine asserts that “true” AI has the ability to simulate human emotions, behavior, and intelligence, though no machine yet has reached that level.[29] In surveys conducted by Hildt (2019), the majority of researchers agreed that current electronic computers are not conscious.[30] However, Hildt clarifies that it may be difficult for humans to determine if a machine has consciousness because it can only be evaluated from a third-person perspective, whereas humans perceive the consciousness of other humans from a first-person perspective.[31]

Should humans be concerned?

While the benefits of unsupervised AI include the ability of these machines to rapidly detect skin cancer, or become personal assistants to humans, Meissner (2020) notes that some programmers fear there is the potential threat that a super-intelligent AI could try to dominate or eliminate the human race.[32] To overcome this potential threat, Meissner, proposes regulations that include keeping robots slow-moving, creating them physically weak, preventing physical proximity to large numbers of other AIs, using a firewall to prevent control of intruders, supervising engineers, creating an on/off switch that is inaccessible to the robot, and creating an overwriting code or “kill-switch”.[33] However, not all researchers in the field share this view of AI as a potential threat.

Dong et al., (2020) believe the future of AI will enhance, rather than replace, the intelligence of humans and predicts they will act as a service provider for human beings, such as when elaborate calculations are  needed. Their research team wants AI to be seen as a tool to assist us in achieving the next stage of civilization, similar to how machines were invented to assist in manufacturing during the industrial revolution, or how agricultural tools were invented to assist in farming. Since AI does not have independent needs, nature, social behavior, or attributes, Dong et al. expect it will be impossible for any AI to become a subject like a human.[34]

Moving Forward

Gardner’s Hype Cycle explains how new technology goes through a phase of overenthusiasm followed by disillusion before an eventual enlightenment of understanding the innovation’s relevance.[35] IBM’s website quotes Lex Fridman as stating, “we are at the peak of inflated expectations, approaching the trough of disillusionment.”[36] However Kelsey Piper, senior writer at Future Perfect, believes otherwise - quoting Greg Brockman (president of OpenAI) as tweeting, “Prediction: 2023 will make 2022 look like a sleepy year for AI advancement & adoption.”[37]

Though the future of consciousness in AIs remains debated, the benefits of developing AI are rapidly becoming apparent. Within the National Institute of Health’s National Human Genome Research Institute (NHGRI), scientists have been using AI to characterize genomic disorders, such as skin conditions, and analyze how communicating this information influences a person's behavior.[38] A new project led by Texas A&M University will use AI to monitor the heating and cooling of buildings through efficient heat pumps to decrease carbon emissions.[39] The field of psychiatry may also soon see its benefits with the recent news that automated analysis of speech patterns may be able to correctly identify depression, psychosis, and suicide risk.[40,41]

While the development of AI is perceived by many as a sign of progress, it has also been associated with negative consequences including rising unemployment, leakage of private information, algorithms that promote bias, and widening of the wealth gap.[42] Dr. Glenn Geher, professor of psychology at the State University of New York, recently stated his concern in Psychology Today where he warns that AI-assisted writing software may reduce thinking abilities and writing skills in future generations.[43] He even illustrates this by writing the majority of his article in the AI software ChatGPT in under 20 seconds.[44]

As scholars continue debating the pros and cons of this rapidly evolving technology, public judgment may come down to people’s individual perception of benefits versus risks. If the robot lawyer is successful in getting its human defendant out of a speeding ticket next month, the public may welcome its assistance with open arms.

Contributed by: Theresa Nair

Editor: Jennifer (Ghahari) Smith, Ph.D.

References

1 Sparkes M. AI legal assistant will help defendant fight a speeding case in court. News Scientist Web site. https://www.newscientist.com/article/2351893-ai-legal-assistant-will-help-defendant-fight-a-speeding-case-in-court/. Updated 2023. Accessed January 7, 2023.

2 Browder J. DoNotPay | The world's first robot lawyer. DoNotPay Web site. https://donotpay.com/about/. Updated 2023. Accessed January 10, 2023.

3 Leffer L. DoNotPay's 'robot lawyer' is gearing up for its first U.S. court case. GIZMODO Web site. https://gizmodo.com/donotpay-speeding-ticket-chatgpt-1849960272. Updated 2023. Accessed January 7, 2023.

4 Dong Y, Hou J, Zhang N, Zhang M. Research on how human intelligence, consciousness, and cognitive computing affect the development of artificial intelligence. Complexity (New York, NY). 2020;2020:1-10. doi:10.1155/2020/1680845

5 What is artificial intelligence (AI)? IBM Web site. https://www.ibm.com/topics/artificial-intelligence. Accessed January 7, 2023.

6 Applications of artificial intelligence across various industries. Forbes Web site. https://www.forbes.com/sites/qai/2023/01/06/applications-of-artificial-intelligence/?sh=410b10923be4. Updated 2023. Accessed January 7, 2023.

7 Hildt E. Artificial intelligence: Does consciousness matter? Frontiers in psychology. 2019;10:1535-1535. doi:10.3389/fpsyg.2019.01535

8 Market Trends. Why 2023 is a breakout year for generative AI? Analytics Insight Web site. https://www.analyticsinsight.net/why-2023-is-a-breakout-year-for-generative-ai/. Updated 2023. Accessed January 7, 2023.

9 Creator N. A picture of artificial intelligence. NightCafe Web site. https://creator.nightcafe.studio/creation/OoeiQbAh08fiJbWuET9t. Updated 2023. Accessed January 10, 2023.

10 Hildt (2019)

11 Ibid.

12 Chrisley R. Philosophical foundations of artificial consciousness. Artificial intelligence in medicine. 2008;44(2):119-137. doi:10.1016/j.artmed.2008.07.011

13 Whang O. ‘Consciousness’ in robots was once taboo. now it’s the last word. The New York Times Web site. https://www.nytimes.com/2023/01/06/science/robots-artificial-intelligence-consciousness.html. Updated 2023. Accessed January 7, 2023.

14 Xie J. An explanation of the relationship between artificial intelligence and human beings from the perspective of consciousness. Cultures of Science (online). 2021;4(3):124-134. doi:10.1177/20966083211056376

15 Kramer B. Mind, symbolism, formalism: Is Leibniz a precursor of artificial Intelligence? . Knowl.Org. 1996;23(2):84-87. https://www.nomos-elibrary.de/10.5771/0943-7444-1996-2-83.pdf.

16 Meissner G. Artificial intelligence: consciousness and conscience. AI & society. 2020;35(1):225-235. doi:10.1007/s00146-019-00880-4

17 Ibid.

18 Banerjee S. A Framework for Designing Compassionate and Ethical Artificial Intelligence and Artificial Consciousness. Interdisciplinary Description of Complex Systems. 2020;18(2-A):85-95. doi:10.7906/indecs.18.2.2

19 Ibid.

20 Xie (2021)

21 Ibid.

22 Zhou Z. Emotional thinking as the foundation of consciousness in artificial intelligence. Cultures of Science (online). 2021;4(3):112-123. doi:10.1177/20966083211052651

23 Xie (2021)

24 Bringsjord S. Meeting Floridi’s challenge to artificial intelligence from the knowledge-game test for self-consciousness. Metaphilosophy. 2010;41(3):292-312. doi:10.1111/j.1467-9973.2010.01639.x

25 Ibid.

26 Pipitone A, Chella A. Robot passes the mirror test by inner speech. Robotics and Autonomous Systems. 2021;144:103838. https://www.sciencedirect.com/science/article/pii/S0921889021001238. doi: 10.1016/j.robot.2021.103838.

27 Whang (2023)

28 Ibid.

29 Forbes (2023)

30 Hildt (2019)

31 Ibid.

32 Meissner (2020)

33 Ibid.

34 Dong et al. (2020)

35 Blosch M, Fenn J. Understanding gartner’s hype cycles. Gartner Web site. https://www.gartner.com/en/documents/3887767. Updated 2018. Accessed January 7, 2023.

36 IBM (2023)

37 Piper K. Think AI was impressive last year? wait until you see what’s coming. Vox Web site. https://www.vox.com/future-perfect/2023/1/4/23538647/artificial-intelligence-chatgpt-openai-google-meta-facial-recognition. Updated 2023. Accessed January 7, 2023.

38 Artificial intelligence and machine learning becoming pervasive at NHGRI and in genomics. NIH | National Human Genome Research Institute Web site. https://www.genome.gov/about-nhgri/Director/genomics-landscape/jan-5-2023-artificial-intelligence-and-machine-learning-becoming-pervasive-at-nhgri-and-in-genomics#ai-ml. Updated 2023. Accessed January 7, 2023.

39 Kulhmann S. Applying artificial intelligence to decarbonize buildings. Texas A&M Today Web site. https://today.tamu.edu/2023/01/06/applying-artificial-intelligence-to-decarbonize-buildings/. Updated 2023. Accessed January 7, 2023.

40 Dikaios K, Rempel S, Dumpala SH, Oore S, Kiefte M, Uher R. Applications of speech analysis in psychiatry. Harvard Review of Psychiatry. 2023;31(1):1-13. https://journals.lww.com/hrpjournal/Abstract/2023/01000/Applications_of_Speech_Analysis_in_Psychiatry.1.aspx. doi: 10.1097/HRP.0000000000000356.

41 DeStafano J. Speech analysis can help measure diagnosis, severity, and onset of mental illness. Neuroscience News Web site. https://neurosciencenews.com/speech-analysis-mental-health-22194/. Updated 2023. Accessed January 10, 2023.

42 Dong et al. (2020)

43 Geher G. ChatGPT, artificial intelligence, and the future of writing. Psychology Today Web site. https://www.psychologytoday.com/us/blog/darwins-subterranean-world/202301/chatgpt-artificial-intelligence-and-the-future-of-writing. Updated 2023. Accessed January 7, 2023.

44 Ibid.