1. The purpose of the IT product Telepathy – the brain-computer interface of Nueralink
The UN General Assembly Resolution of 22.03.2024 No. A/78/L.49 “Harnessing the Potential of Safe, Secure and Trustworthy Artificial Intelligence Systems for Sustainable Development” [1] confirms (paragraph 5) that"human rights and fundamental freedoms must be respected, protected and encouraged throughout the life cycle of artificial intelligence systems"and to all Member States of the United Nations and, where appropriate, In such cases, other interested parties should"refrain from using or stop using artificial intelligence systems that cannot be operated in accordance with international law or that create undue risks to the enjoyment of human rights, particularly for those in vulnerable situations,”at the same time"The rights that people have offline must also be protected online, including throughout the life cycle of artificial intelligence systems".
The artificial intelligence (“AI”) systems currently being created, tested and deployed are covered by the broader concept of information technology product (“IT product”).
In turn, the author of the term artificial intelligence, Stanford University professor John McCarthy, defined intelligence as"computing part the ability to achieve goals in the world", and AI is like"science and technology of creation intelligent machines, especially intelligent computer programs". At the same time, the same author predicts or laments that"Once this works, no one else will will not call it AI."[20]
It is logical to assume that we will return to the use of a general and established understanding and definition of an IT product, which is usually understood as a comprehensive software and hardware solution in the field of information technology, designed to perform certain functions.
As we know, the creation and functioning of any IT product (including AI tools) is based on mathematical, computational procedures, classical filters and matrix equations.
An IT product may include both software for various purposes and hardware components, as well as their combinations, including various types of technical means for processing and transmitting information (including software interfaces), which are circulated on the information market and are distinguished by numerous features both at the stages of development, production, and circulation. [13, 14]
In general, the purpose of an IT product is understood as satisfying the needs of users in the field of information technology. At the same time, the classification of IT products can be varied and depends on many parameters, in particular, such as functionality, scope of application, architecture and technology.
The purpose of the brain-computer interface is conceptually defined by its creators as follows:"To regain independence and improve life, we created a brain-computer interface (BCI) that provides fast and reliable computer control and prioritizes ease of use."[2-6].
In other words, such innovative products have the primary goal of improving the quality of life of people suffering from limb disabilities or other paralytic conditions. [2-6]
In addition to achieving the goals of providing solutions for various neurological disorders, Telepathy IT product has the following practical potential [2-6]:
1. Improving mental health: Neuralink technology has potential applications in the treatment of mental illnesses. It may be possible to develop interventions for conditions such as depression, anxiety, or obsessive-compulsive disorder;
2. Cognitive enhancement: The company's founder, Elon Musk, suggests that Neuralink could potentially improve human cognitive abilities such as memory, learning speed, or information processing;
3. Human-Machine Integration: Neuralink can facilitate seamless integration between people and technology, including improved communication between people, better control of external devices, or more immersive virtual reality experiences;
4. Brain Research and Understanding: The technology developed by Neuralink could provide valuable insights into how the brain actually functions, which could have significant implications for neuroscience and cognitive science in general.
Nueralink is currently developing a spinal cord chip that is planned to be used specifically through the Telepathy IT product – a brain-computer interface (BCI).
In the future, Neuralink’s brain-computer interface could be aimed at restoring vision to the blind and other people with visual impairments. Neuralink’s brain chip could stimulate parts of the visual cortex and perhaps even restore vision when the connection between the optic nerve and the brain is lost. The product is planned to be called “Blindsight.”[2-6]\
2. Description of the life cycle of an IT product
Established terminology and a common recommended framework for the life cycle processes of IT products that can be used as a reference in the software industry, are reflected in a number of standardization documents at both international and national levels. [13-14]
So, as we know, the life cycle of an IT product is defined as development of a system, product, service, project or other man-made objects, from the concept stage to the end of use. In this case, the IT product (software product) is understood as a set of computer programs, procedures and possibly related documentation and data. [13-14]
The life process of an IT product is described by a life cycle model consisting of stages. Models are commonly used to represent the entire life cycle from conception to termination of use or to represent only a part of the life cycle of a specific project.
The life cycle model is reduced to the description of a sequence of stages that may overlap and/or repeat cyclically in accordance with the scope, size, complexity of the IT product, the need for changes and capabilities. Each stage is described in definitions of its purpose and outputs.
The general approach to defining the life cycle of an IT product is also maintained when describing the life cycle of artificial intelligence tools (AI tools).
For example, in the current UK regulation we find the following definition:"The AI life cycle: all events and processes associated with the lifespan of an AI system from creation to retirement, including its design, research, training, development, implementation, integration, operation, maintenance, sale, use and management."[15]
In other words, the life cycle of AI products is iterative [16] – just like the life cycle of an IT product of any vendor operating in the open market and maintaining feedback with the product’s users (customers and/or consumers).
In turn, the consolidated approach of the European Union to defining the life cycle of AI is aimed at containing the uncontrolled, socially dangerous and unfair use of AI tools and, for these purposes, comes down to the following model (stages) of the life cycle of AI [16]:
1. Design in line with requirements – at this stage, it is important for IT product developers/suppliers to ensure that the AI systems operate stably as intended and are in working order in accordance with regulatory standards;
2. Conformity assessment – here a preliminary assessment of the conformity of the IT product is performed;
3. Post-market monitoring – a phase in which developers/vendors should actively and systematically collect, document and analyze relevant data on the reliability, performance and safety of AI systems throughout their service life, and assess and confirm the continued compliance of AI systems with regulatory standards;
4. Incident report system – the goal and output at this stage is timely notification by developers/suppliers about serious incidents, as well as about malfunctions leading to violations fundamental rights (as a basis for investigations carried out by competent authorities);
5. New conformity assessment – a cyclical assessment of conformity in the event of a significant change in the IT product (a change in its intended
purpose or a change affecting the compliance of the AI system with regulatory standards) by developers, suppliers or any third party, including when the changes go beyond the “predefined range” specified by the developer/supplier for continuously learning artificial intelligence systems.
Thus, we can talk about historical continuity between the generally accepted approaches to the continuous cycle of product testing (including IT products), formulated in both international and national regulations. According to the meaning of these regulations, the object of testing an IT product (AI product) is the IT product as a whole.
3. Description of the life cycle of the IT product Telepathy – the brain-computer interface (BCI) of Nueralink
Based on the presented understanding of the definition of an information product, as well as the methods and standards related to the organization of the life cycle of information products, it is permissible to present the following description of the life cycle of the IT product of artificial intelligence Telepathy - the brain-computer interface (BCI) of the Nueralink company:
1. Design in line with requirements – Neuralink started developing the Telepathy project in 2016;
2. Conformity assessment – over the next almost seven years, only animals were implanted: a monkey was taught to move a cursor with the power of thought, and pigs were given impulses to perform movements. In the fall of 2023, the US Food and Drug Administration granted Neuralink permission to conduct human trials as part of the Telepathy project;
3. Post-market monitoring – in the fall of 2023, Neuralink began selecting candidates for brain chip implantation, and on January 28, 2024, the first person received an implant from Neuralink, the patient began to successfully recover and the promising formation of neural spikes in his brain (activated using the Telepathy IT product - a brain-computer interface);
4. Incident report system – informing competent authorities about incidents and/or malfunctions requiring investigation, which have not occurred at the moment and the tests are proceeding successfully;
5. New conformity assessment – the first practical clinical applications of the IT product of artificial intelligence Telepathy - the brain-computer interface (BCI) of Nueralink are planned within five to seven years, which will allow the product to be improved and ensure its compliance with the changing purposes of its application.
Referring to a brief description of the iterative life cycle of the Telepathy artificial intelligence IT product (Nueralink's brain-computer interface (BCI)), the following can be noted:
1. The Pre-Design stage has been successfully completed - the project planning for the creation of an IT product, the problem specification have been completed, the research background has been formed, and the data from animal tests necessary for designing a working model of the IT product have been collected;
2. The Design and Development stage has been successfully completed - requirements and collected data have been analyzed, a working model of the IT product has been designed, and permission for clinical trials on humans has been obtained;
3. Deployment stage – currently underway. The first working model of the IT product Telepathy (brain-computer interface (BCI) by Nueralink) has been released and is in use. Within a year, it is planned to put into operation 11 more such interfaces, and within five to seven years – mass application of the brain-computer interface (BCI) for at least three categories of people with disabilities - those with limb disabilities, spinal cord injuries, and visual injuries/pathologies/functional disorders.
4. Conclusion
In summary, it should be noted that the expert community and business environment recognize and understand the timeliness of political efforts aimed at creating regulatory standards applicable to the life cycle of artificial intelligence IT products.
ACM members Weiyue Wu and Shaoshan Liu summed up the regulatory dilemma in the field of artificial intelligence:
“Without a properly regulated environment, AI companies will not be able to thrive and transform their latest technologies into successful products. When regulations are lacking, players in the field can go crazy and gamble both the company’s future and public safety, which jeopardizes technological progress in the long term. When regulations are too stringent, budget-strapped AI companies, especially startups, will suffer irreparable financial losses due to repetitive filings with multiple offices. The key to resolving the problems is a clear message that the potential benefits outweigh the associated risks.” [18]
The current categorization of risks of implementing artificial intelligence IT products has been fundamentally formulated by expert members of the Association for Computing Machinery (ACM) and is reflected in the latest legislation, in particular, in the European Union Artificial Intelligence Act, as well as in US presidential decrees and acts of the United Nations.
Table 1. Artificial Intelligence Risk Levels and Regulatory Measures
Risk level
|
Risk content
|
Regulatory measures |
Unacceptable risk |
Artificial intelligence systems are considered a clear threat to the security of existence and legally protected rights of people, for example, those systems that use subconscious manipulation or social rating |
Prohibition developments/ use |
High risk |
Artificial intelligence systems that could cause harm if they malfunction or are misused, such as in education, employment, law enforcement, immigration, product safety, critical infrastructure |
Regulation developments/ use |
Limited risk |
AI systems at risk of manipulation or deception, where people must be informed that they are interacting with AI, such as a chatbot or emotion recognition tool
|
Transparency use |
Minimal risk |
Most AI systems currently in use, such as video games or spam filters |
Without imputation obligations for developers/ to suppliers
|
We see governments setting limits on AI - many countries/regions are discussing or trying to implement regulations on the training and use of artificial intelligence.
The above demonstrates the importance of not only ensuring exponential growth in the field of creation and development of IT products of artificial intelligence, but also the importance parallel funding and implementation of research in the field of neuroethics to ensure the protection of humanity from its own activity and the products of its intellectual creativity and progress.
It seems important to discuss neuroethics issues now, at the early stages of the development and implementation of AI tools, especially in such a sensitive area as medicine.
The regulation of neuroethics issues in sensitive areas is also supported by the recognition by one of the fathers of the Internet, Vinton Gray "Vint" Cerf, that "when we invented the Internet... we did not think about the consequences." [21]
Our generation has a chance to think and act more responsibly and insightfully.
5. Bibliography
1. UN General Assembly Resolution of 22.03.2024 No. A/78/L.49 “Harnessing the Potential of Safe, Secure and Trustworthy Artificial Intelligence Systems for Sustainable Development” / Official website of the United Nations (Original: English) – URL: https://documents.un.org/doc/undoc/ltd/n24/065/92/pdf/n2406592.pdf (date of access 29.09.2024);
2. A Seamless BCI Experience – Neuralink.com - URL: https://neuralink.com/#experience (date of access - 24.03.2024);
3. Neuralink Live Update - March 2024 / @neuralink channel - URL: https:// www.youtube.com/watch?v=ZzNHxC96rDE (accessed on 22.03.2024);
4. Neuralink Begins First Human Experiments! / @TheTeslaSpace channel - URL: https://www.youtube.com/watch?v=CxBbw8eT624 (date of access - 22.03.2024);
5. Inside Neuralink's Brain Chip: How Does Elon Musk's New Tech Vision Work / Outlook Business Desk // Outlook India - URL: https://business.outlookindia.com/news/inside-neuralinksbrain-chip-how-does-elon- musks-new-tech-vision-work (accessed April 13, 2024);
6. A Non touch interactions/ Prof. R. Gayathri // Bannari Amman Institute of Technology - URL: https://www.bitsathy.ac.in/telepathy-a-non-touch-interactions/ (date of access - 08.04.2024);
10. Artificial intelligence is just a new computational procedure / Academician of the Russian Academy of Sciences Igor Sokolov // Russian Academy of Sciences - URL: http://www.ras.ru/news/shownews.aspx?id=b8da5534-67fd-402e-9cad-876b921cbc5a&print=1 (date of access – 16.04.2024);
11. Sci-Fi to reality: Elon Musk's brain chip implanted in human brain by Neuralink / ET Edge Insights by Times Group - URL: https://etinsights.et-edge.com/sci-fi-to-reality-elon-musks-brainchip-implanted-in-human-brain-by-neuralink/ (date of access – 13/04/2024);
12. Elon Musk Teases First Neuralink Products After Company Implants First Brain Chip In Human / R. Hart // Reuters - URL: https://www.reuters.com/technology/neuralink-implants-brainchip-first-human-musk-says-2024-01-29/#:~:text=The%20first%20product%20from%20Neuralink,Reuters%20request%20for%20further %20details (date of access – 13.04.2024);
13. ISO/IEC/IEEE 12207:2017 Systems and software engineering - Software life cycle processes / URL: https://www.iso.org/obp/ui/ru/#iso:std:iso-iec-ieee:12207 :ed-1:v1:en - Online Browsing Platform, OBP (accessed March 23, 2024);
14. Guide to the software engineering body of knowledge (SWEBOK V3.0) / – Editors Pierre Bourque, École de technologie supérieure (ÉTS), Richard E. (Dick) Fairley, Software and Systems Engineering Associates (S2EA) - B.: IEEE, 2014. – 335 p.;
15. A pro-innovation approach to AI regulation: government response / GOV.UK - URL: https:// www.gov.uk/government/consultations/ai-regulation-a-pro-innovation-approach-olicyproposals (date of access 04/13/2024);
16. A European Strategy for Artificial Intelligence CEPS webinar / L.Sioli // – B.: European Commission, 2021 – 21 p.;
17. Scherer, MU Regulating artificial intelligence systems: Risks, challenges, competencies, and strategies / Harvard Journal of Law and Technology - URL:https://bit.ly/3YdCqSY (date of access 16.04.2024);
18. Dilemma of the Artificial Intelligence Regulatory Landscape / Weiyue Wu, Shaoshan Liu // Communications of the ACM, 2023 – URL: https://cacm.acm.org/opinion/dilemma-of-theartificial-intelligence-regulatory-landscape/ (date of access 17.04.2024);
19. Towards a Standard for Identifying and Managing Bias in Artificial Intelligence NIST Special Publication 1270/ R.Schwartz, A. Vassilev, K.Greene, L.Perine, A.Burt, P.Hall // US Department of Commerce, 2022 – URL: https://doi.org/10.6028/NIST.SP.1270(date of access 17.04.2024);
20. What is Artificial Intelligence? / Prof. John McCarthy // John McCarty's Home Page, 2007 – URL: http://www-formal.stanford.edu/jmc/whatisai.pdf(date accessed 17.04.2024).
21. A Brief But Spectacular - Taking on the future of the internet - Vint Cerf / The PBS NewsHour, 2024 - URL: https://www.pbs.org/newshour/amp/show/a-brief-but-spectacular-take-on-the-futureof-the-internet (date accessed 17.04.2024).
