Is NVIDIA’s Omniverse ethical?

Professor Edmond C. Prakash
University for the Creative Arts, UK
eprakash@gmail.com

Professor Andrés A. Navarro-Newball
Pontificia Universidad Javeriana Cali, Colombia
anavarro@javerianacali.edu.co

1. Introduction.

“As the concept of the Metaverse grows in popularity, many scholars and developers begin to focus on the Metaverse’s ethics and core, (Yang et al., 2023).” Here (Yang et al., 2023), the idea is to combine development practices, interdisciplinary methods, to develop an ethical and responsible Metaverse utilising meaningful human control to ensure human decision-making power, privacy, ethics and the rights and interests of fairness. Thus, the idea behind the so-called human centric Metaverse (Yang et al., 2023) is to have a human core so that creations focus on human services and always supports the user experience. Bill Gates stated several years ago (1995) that technology should be the servant to us and not the master. As Yang et al. (2023) highlights: “although digital ethics is not dissimilar to traditional ethics, there are risks associated with large-scale unintentional or intentional unethical behaviour. With the uncontrolled technological development of the Metaverse, previous studies mostly highlighted the co- existing opportunities and threats for humanity.”

As a long-time passionate professors and developers in Computer Graphics and its applications, we decided to provide our opinion on the NVIDIA’s Omniverse (NVIDIA Corporation, 2023), which recently had an update. We used Microsoft’s Copilot from the Bing search engine as a Generative AI tool to help usunderstand the challenges involved. The references cited in this brief introduction were read, understood, and processed by us. All other sources were found by Bing. The content in the next few sections was generated using Bing but carefully curated by us (we put an * next to the title of these sections). Section 7, where we discuss son trends on regulations, was written with a mixture of both (50% – Human will), but also carefully curated. Except from one final remark about power consumption, the final discussion is from our own authorship.

2. Some relevant definitions (*)

Generative AI (GAI) is a branch of artificial intelligence that can create original content by learning patterns from existing data. The Metaverse refers to a collective virtual shared space created by the convergence of physical and digital realities. It encompasses various interconnected virtual worlds, social platforms, and immersive experiences. Primarily virtual and digital, the Metaverse focuses on creating immersive experiences within purely synthetic environments. Thus, it focuses on how users interact within virtual environments. The Omniverse extends beyond the Metaverse. It encompasses not only virtual spaces but also includes the entire interconnected fabric of digital, physical, and hybrid environments. This broader scope includes real-world data, simulations, and interactions. It includes digital twins, simulations, and data-driven models of physical objects, cities, and ecosystems. Thus, it expands to include how humans interact with both digital and physical components. This includes ethical behaviour in augmented reality (AR), Internet of Things (IoT), and smart cities.

3. So, what are the ethical challenges? (*)

All of them, GAI, the metaverse and the omniverse have tremendous potential for innovation and creativity, but they also pose some major ethical concerns.

The metaverse poses concerns about data privacy, surveillance, and user tracking within virtual spaces. Aso, there are challenges related to digital identity, avatars, and anonymity. Additionally, there are issues balancing freedom of expression with preventing harmful or illegal content. Finally, there are discussions around virtual property rights, intellectual property, and decentralised governance.

Meanwhile, the omniverse poses concerns on how digital representations of physical objects impact the real world (e.g., autonomous vehicles, smart cities). Also, there are considerations regarding energy consumption, resource usage, and sustainability. Additionally, there are challenges ensuring seamless communication between virtual and physical systems. Discussions extend to ownership of real-world assets (e.g., buildings, infrastructure) represented digitally and governed by complex systems.

On the other hand, GAI can generate realistic and personalised content, such as images, videos, texts, and voices, based on user data. However, this may infringe on the privacy and consent of the users and the subjects of the content. For example, GAI can create deepfakes, which are synthetic media that manipulate the appearance or speech of real people. Deepfakes can be used for malicious purposes, such as impersonation, fraud, or defamation. Moreover, GAI can also generate content that may reveal sensitive
or personal information about the users or the subjects, such as their preferences, behaviours, or identities. Users may not have full control or awareness of how their data is collected, used, or shared by GAI applications in the metaverse.

GAI relies on large amounts of data to learn and generate content. However, the data may be biased, incomplete, or inaccurate, which can affect the quality and fairness of the content. For example, GAI can generate content that reflects or amplifies existing stereotypes, prejudices, or discrimination against certain groups of people. Bias can also affect the accessibility and inclusivity of GAI applications in the metaverse, such as the representation and diversity of avatars, languages, and cultures. Moreover, bias can also affect the outcomes and decisions of GAI applications, such as the recommendations, predictions, or evaluations that they provide.

GAI can generate content that is indistinguishable from human-generated content, which can raise questions about the authenticity and credibility of the content. For example, GAI can generate content that is misleading, deceptive, or false, such as fake news, propaganda, or misinformation. Users may not be able to verify the source, origin, or intention of the content, or distinguish between real and synthetic content. Moreover, GAI can also generate content that is complex, novel, or unexpected, which can challenge the explainability and interpretability of the content. Users may not be able to understand how or why the content was generated, or what the implications or consequences of the content are. Furthermore, GAI can also generate content that is autonomous, dynamic, or interactive, which can blur the boundaries and responsibilities between human and machine agency. Users may not be able to attribute or hold accountable the creators, developers, or users of the content for the impacts or harm that the content may cause.

GAI can generate content that is influenced by or affects the ethical values and norms of the users and society. For example, GAI can generate content that is offensive, harmful, or unethical, such as hate speech, violence, or pornography. Users may not be able to regulate or moderate the content or protect themselves or others from the content. Moreover, GAI can also generate content that is expressive, creative, or artistic, such as music, poetry, or stories. Users may not be able to appreciate or respect the content or acknowledge the intellectual property or moral rights of the content. Furthermore, GAI can also generate content that is immersive, engaging, or addictive, such as games, simulations, or experiences. Users may not be able to balance or integrate the content with their physical, social, or psychological well-being.

4. What about NVIDIA’s Omniverse? (*)

As an example, NVIDIA, a prominent technology company, places significant emphasis on ethical considerations across its various initiatives, including the Omniverse. NVIDIA OmniverseTM is a powerful computing platform that enables individuals and teams to develop Universal Scene Description (USD)- based 3D workflows and applications. Omniverse unifies 3D data from various sources, allowing creators to work seamlessly across different applications. You can sync your favourite creative tools to Omniverse and USD, providing a unified view of your 3D data. Creators can use Omniverse to create in 3D faster than ever. Connect your creative apps and work with your data efficiently. Developers can build custom extensions, tools, and microservices to enhance 3D workflows. Omniverse supports low-code development augmented by AI. Omniverse extends beyond virtual spaces. It includes digital twins—digital representations of physical objects, cities, and ecosystems. Enterprises can break 3D data silos, unify large teams, and enable advanced 3D and simulation workflows. NVIDIA has recently released a major Omniverse upgrade with Generative AI and OpenUSD.

NVIDIA has established a comprehensive set of policies and guidelines that guide its business practices. These include the Code of Conduct, which outlines ethical behaviour expectations for employees and stakeholders. The company also maintains a Speak Up Line for confidentially reporting concerns related to ethics and compliance. Additionally, NVIDIA is actively addressing climate change within the Omniverse. They are developing a digital Earth called E2 (Earth 2), which serves as a supercomputer model to predict regional impacts of climate change up to 30 years into the future. This digital twin of Earth aims to accurately model the planet’s climate, contributing to climate research and sustainability efforts. The Omniverse platform strives for universal interoperability across different applications and vendors. It facilitates real-time scene updates and adheres to open standards and protocols. By promoting openness and collaboration, NVIDIA encourages ethical practices in the development and use of virtual environments. NVIDIA engages in conversations about creating trustworthy AI. Their corporate responsibility report includes discussions on societal impacts, including the challenges of building ethical AI systems.

Microsoft Azure will host two cloud offerings from NVIDIA: NVIDIA OmniverseTM Cloud, a platform-as-a- service for building and operating 3D industrial metaverse applications; and NVIDIA DGXTM Cloud, an AI supercomputing service for training advanced models for generative AI and other applications. Additionally, the companies are connecting Microsoft 365 applications with NVIDIA Omniverse, enabling real-time 3D collaboration across different productivity and creative tools. This collaboration aims to accelerate enterprises’ ability to digitalise their operations, engage in the industrial metaverse, and train advanced models for generative AI and other applications.

5. What are the ethical concerns? (*)

There are some ethical concerns regarding the collaboration between NVIDIA and Microsoft on the Omniverse and AI supercomputing. NVIDIA’s chips are widely used for various AI applications, such as self- driving cars and image recognition. However, these applications may suffer from AI bias, which can lead to unfair or inaccurate outcomes for certain groups of people. Moreover, there are diversity issues in the AI industry, which may affect the representation and inclusion of different perspectives and values in AI development and deployment. The Omniverse and AI supercomputing platforms rely on large amounts of data, which may pose risks to data privacy and security. Users may not have full control or awareness of how their data is collected, stored, processed, and shared by these platforms. Data breaches or misuse may also expose sensitive or personal information to unauthorised parties. The Omniverse and AI supercomputing platforms enable advanced and complex 3D and AI applications, which may have significant impacts on society and the environment. However, there may not be clear or consistent ethical standards and governance mechanisms to regulate these applications and ensure they are aligned with human values and rights. Moreover, there may be challenges in ensuring accountability and transparency for the decisions and actions of these platforms and their users.

6. How are these concerns addressed? (*)

These concerns may be addressed by adopting responsible and ethical AI practices, such as ensuring data quality and fairness, protecting data privacy and security, establishing ethical guidelines and frameworks, and engaging in stakeholder collaboration and consultation. Indeed, NVIDIA and Microsoft are aware of the ethical concerns of Omniverse and AI supercomputing, and they are taking steps to address them. They are testing their datasets for potential bias, adding warning labels to synthetic media, and collaborating with affinity groups to promote diversity and inclusion in their workforce and products. They are following industry standards and best practices, such as encryption, authentication, and compliance. They are also providing users with control and transparency over their data collection, usage, and sharing. They are participating in industry initiatives, such as the Partnership on AI and the Responsible AI Institute, to establish ethical principles and norms. They are also consulting with stakeholders, such as customers, regulators, and experts, to ensure accountability and responsibility for their AI impacts.

7. Fire is a good servant but a bad master. Where are the AI guardrails, AI escape routes, AI extinguishers, AI sand and water?

7.1. AI Helper vs Master

At a recent event on “AI’s Impact on Building Bonds’ on LinkedIn Live” hosted by Dr. Pramila Thapa (Thapa, 2024), the conversation between Michael Rada (Rada, 2024) and Matthew Young (Young, 2024) brought in new insight on how Fire can be related to AI. The analogy “fire is a good servant but a bad master” can indeed be applied to AI. Here’s how it can be interpreted in the context of AI:

Good Helper: AI can be an incredibly powerful tool when used responsibly and ethically. It can assist humans in various tasks, automate processes, and enhance productivity across numerous domains, from healthcare to transportation to finance. Like a good helper or tool, AI can amplify our capabilities and makeour lives easier.

Bad Master: However, when AI is misused or deployed without proper oversight, it can lead to unintended consequences and even harm. Without appropriate regulations and ethical guidelines, AI systems can perpetuate biases, invade privacy, and exacerbate inequality. Moreover, AI systems can malfunction or be manipulated, leading to unpredictable outcomes or even catastrophic failures. In this sense, AI can become a “bad master” if not carefully controlled.

7.2. AI Guardrails

Establishing ethical guidelines, regulatory frameworks, and promoting transparency are essential components of implementing effective AI guardrails to prioritize human well-being and address concerns regarding privacy, accountability, and fairness in AI systems.

Ethical Guidelines: Establishing clear ethical principles for AI development and deployment can help ensure that AI systems prioritize human well-being and adhere to moral standards.

Regulatory Frameworks: Governments and organizations can implement regulations to govern the use of AI, addressing concerns such as data privacy, algorithmic transparency, and accountability.

Transparency and Explainability: Promoting transparency in AI systems and ensuring that they are explainable can help mitigate the risks of bias, discrimination, and unintended consequences.

Robustness and Security: Developing AI systems that are resilient to adversarial attacks and cybersecurity threats can safeguard against potential misuse or manipulation.

7.3. AI Escape Routes

To safeguard against potential risks and ensure responsible AI deployment, strategies such as maintaining human oversight, conducting algorithmic audits, enabling continuous learning, and incorporating fallback mechanisms are essential for mitigating harm and promoting ethical use of AI.

Human Oversight: Maintaining human oversight and intervention in AI systems can help mitigate risks and prevent autonomous decision-making that could lead to undesirable outcomes.

Algorithmic Auditing: Regularly auditing AI algorithms and systems can help identify and address biases, errors, and vulnerabilities, ensuring that they operate in accordance with ethical and regulatory standards.

Continuous Learning and Adaptation: Implementing mechanisms for AI systems to learn and adapt over time can enable them to respond to new challenges and changing circumstances while minimizing the potential for harm.

Fallback Mechanisms: Incorporating fallback mechanisms or fail-safes in AI systems can provide a means of reverting to a safe state or alternative course of action in the event of unexpected behavior or failure.

By implementing these guardrails and escape routes, we can strive to harness the potential of AI as a beneficial helper or tool while mitigating the risks associated with its misuse or unintended consequences.

7.4.Global Initiatives and some key AI guardrails

Large scale consultations (UK AI Consult, 2024), research and development to support regulation has just begun as highlighted by the numerous posts by Peng (Peng, 2024). The AI Safety Institutes at the UN (UNAIAB (2024), UK (UKASI, 2024; UKASI2, 2024, UK AISS, 2023), US (USAISI, 2023), ASEAN (ASEAN AI, 2024), etc, aim to establish guardrails for responsible AI. ASEAN effort provides recommendations on national level and regional level initiatives that governments in the region can consider implementing to design, develop, and deploy AI systems responsibly. Other international efforts such as the ISO AI standards, includes ISO/IEC 22989vdefines AI terminology; ISO/IEC 23053 provides a framework for AI and machine learning; and ISO/IEC 23894 provides guidelines for AI-related risk management. ISO/IEC 42001 provides overarching governance.

8. Discussion

We believe GAI has come to stay and we need to learn how to live with it and make the most of it. We are still learning, however, according to what we found and to what we got using Bing, the path to maintain an ethical use of generative AI and particularly an ethical use of the NVIDIA’s omniverse includes:

Using open standards.
Maintaining stakeholders informed.
Working and discussing with regulators.
Using these technologies for applications that matter.
Staying human centric.
Protecting privacy and truthfulness.
Favouring accessibility and inclusion.
Recognising authorship and respecting intellectual property.

GAI helped us provide a better and faster opinion. We will continue using GAI technology to be more productive, always staying as the ones who master technology and not the opposite way, as Bill Gates (1995) used to highlight. We believe most concerns can be addressed with will and hard work. We will probably make mistakes, but also, we will have the chance to amend them. Collaborations like the Global AI & Humanity Centre will help us get there. Of course, more formal research should use GAI as a starting point and have more human intervention.

Also, we want to explore further NVIDIA’s Omniverse. However, we still have a major concern: the cool machines and GPUs consume a lot of energy. One of us have just bought one great Surface Studio Laptop with a fantastic RTX 4060 to explore the Omniverse. We feel it is cool, but at the same time, we feel we are not contributing to lower energy consumption. NVIDIA’s EARTH 2 project helps us understand how we affect the environment but does not diminish power consumption. That said, and again, according to Bing’s Copilot: “The NVIDIA GeForce RTX 4060 draws power from a single 12-pin power connector and has a maximum power draw of 115 watts. This power consumption is significantly lower than the NVIDIA GeForce RTX 3060, which has a TDP of 170 watts;” thus, we are improving.

9. Conclusion

Is the NVIDIA’s omniverse ethical? Well, we think it is. NVIDIA and its partners are following several ethical guidelines. Still, there are some ethical glitches we are aware of, that we should continue working on. It is up to us as humans to keep and eye on how we use GAI, and how we observe what big companies are doing with it.

References

ASEAN AI (2024) ASEAN Guide on AI Governance and Ethics Available at: https://asean.org/book/asean- guide-on-ai-governance-and-ethics/ (Accessed: 16 February 2024).

Gates, B. (1995). The Road Ahead (2nd ed.). New York, NY: Viking.
NVIDIA Corporation. (2023, August 8). NVIDIA Releases Major Omniverse Upgrade With Generative AI
and OpenUSD. Retrieved from NVIDIA Newsroom
Peng (2024), Yonghong Peng, https://www.linkedin.com/in/yonghpeng/

Rada (2024) Michael Rada, Human & Industry 5.0 Founder, Pilsen, Plzeň, Czechia
https://www.linkedin.com/in/michaelrada/

Thapa (2024) Pramila Thapa https://www.linkedin.com/in/pramila-thapa-phd-59616516b/

UK AI Consult (2024) A pro-innovation approach to AI Regulation: Government response, GOV.UK.

Available at: https://www.gov.uk/government/consultations/ai-regulation-a-pro-innovation-
approach-policy-proposals/outcome/a-pro-innovation-approach-to-ai-regulation-government-
response (Accessed: 16 February 2024).

UKAI Leadership (2024) UK signals step change for regulators to strengthen AI leadership, GOV.UK.

Available at: https://www.gov.uk/government/news/uk-signals-step-change-for-regulators-to-
strengthen-ai-leadership (Accessed: 16 February 2024).

UKAI Productivity (2024) Business and tech heavyweights to boost productivity through AI, GOV.UK.

Available at: https://www.gov.uk/government/news/business-and-tech-heavyweights-to-boost-
productivity-through-ai (Accessed: 16 February 2024).

UKAISI (2024) Ai Safety Institute: Overview, Department for Science, Information and Technology,

GOV.UK. Available at: https://www.gov.uk/government/publications/ai-safety-institute-
overview (Accessed: 16 February 2024).

UKAISI2 (2024) Introducing the AI Safety Institute, GOV.UK. Available at:

https://www.gov.uk/government/publications/ai-safety-institute-overview/introducing-the-ai-
safety-institute (Accessed: 16 February 2024).

UK-AISS (2023) AI Safety Summit (2023) AISS 2023. Available at: https://www.aisafetysummit.gov.uk/
(Accessed: 16 February 2024).

UNAIAB (2024) Governing AI for humanity. Available at:
https://www.un.org/sites/un2.un.org/files/ai_advisory_body_interim_report.pdf (Accessed: 16
February 2024).

USAISI (2024) U.S. Artificial Intelligence Safety Institute (2024) NIST. Available at:
https://www.nist.gov/artificial-intelligence/artificial-intelligence-safety-institute (Accessed: 16
February 2024).

Yang, R., Li, L., Gan, W., Chen, Z., & Qi, Z. (2023). The Human-Centric Metaverse: A Survey. In Companion
Proceedings of the ACM Web Conference 2023 (WWW ’23 Companion) (pp. 1296–1306).
Association for Computing Machinery. https://doi.org/10.1145/3543873.3587593

Young (2024) Matthew Young, Founder & Director – studio3eight
https://www.linkedin.com/in/matthewyoung2/

Source(s) used by Microsoft’s Copilot using Bing (*)

Ahuja, V. (2024, February 14). Ethical Considerations for Generative AI. Forbes.

https://www.forbes.com/sites/forbestechcouncil/2024/02/14/ethical-considerations-for-
generative-ai/

Benjamins, R., Viñuela, Y. R., & Alonso, C. (2023). Social and ethical challenges of the metaverse. AI And
Ethics, 3(3), 689–697. https://doi.org/10.1007/s43681-023-00278-5
Clinton, D. (2023, December 18). Nvidia faces concerns over AI bias and diversity issues. Cryptopolitan.
https://www.cryptopolitan.com/nvidia-faces-concerns-over-ai-bias/

Ethics in the metaverse. (n.d.). Frontiers. https://www.frontiersin.org/research-topics/56357/ethics-in-
the-metaverse

Feltes, I. V. (2022, January 31). Preparing businesses for the metaverse and omniverse era. Forbes.

https://www.forbes.com/sites/forbesbusinesscouncil/2022/01/31/preparing-businesses-for-
the-metaverse-and-omniverse-era/

Fernandez, C. B. (2022, March 25). Life, the Metaverse and Everything: An overview of privacy, ethics,
and governance in metaverse. arXiv.org. https://arxiv.org/abs/2204.01480
Gralla, P. (2023, May 23). Ethics — what ethics? For Microsoft, it’s full speed ahead on AI.

Computerworld. https://www.computerworld.com/article/3697014/ethics-what-ethics-for-
microsoft-its-full-speed-ahead-on-ai.html

healthy Pro-teen. (2023, June 28). NvIdia GeForce RTX 4060 Review: Truly Mainstream at $299. Tom’s

Hardware. https://www.tomshardware.com/reviews/nvidia-geforce-rtx-4060-review-asus-
dual/8

Hope, B. (2022, January 18). How Nvidia is fighting climate change in the Omniverse. Sustainability
Magazine. https://sustainabilitymag.com/esg/how-nvidia-fighting-climate-change-omniverse

Kaburu, G. (2023, September 1). Advocating ethical use of Nvidia’s AI chips and DeepMind Co-Founder’s

proposal. Cryptopolitan. https://www.cryptopolitan.com/advocating-ethical-use-of-nvidias-ai-
chips/

NVIDIA and Microsoft to bring the industrial metaverse and AI to hundreds of millions of enterprise users

via Azure Cloud. (n.d.). NVIDIA Newsroom. https://nvidianews.nvidia.com/news/nvidia-and-
microsoft-to-bring-the-industrial-metaverse-and-ai-to-hundreds-of-millions-of-enterprise-users-
via-azure-cloud

NVIDIA GeForce RTX 4060 Specs. (n.d.). TechPowerUp. https://www.techpowerup.com/gpu-
specs/geforce-rtx-4060.c4107

NVIDIA Omniverse. (n.d.). NVIDIA. https://www.nvidia.com/en-us/omniverse/
NVIDIA Omniverse Cloud. (n.d.). NVIDIA. https://www.nvidia.com/en-us/omniverse/cloud/
Nvidia RTX 4060 graphics card power consumption test: Game Power … (n.d.).

https://www.techgoing.com/nvidia-rtx-4060-graphics-card-power-consumption-test-game-
power-consumption-126w/

NVIDIA teams with Microsoft to build massive cloud AI computer. (n.d.). NVIDIA Newsroom.
https://nvidianews.nvidia.com/news/nvidia-microsoft-accelerate-cloud-enterprise-ai
O’Donnell, B. (2021, November 3). Microsoft and Nvidia are working on their own more practical

metaverse. TechSpot. https://www.techspot.com/news/92064-microsoft-nvidia-working-their-
own-more-practical-metaverse.html

Omniverse Structure — Omniverse Platform latest documentation. (n.d.).
https://docs.omniverse.nvidia.com/platform/latest/structure.html
Read about NVIDIA’s code of conduct, trademark and social media guidelines, and environmental policy.
(n.d.). NVIDIA. https://www.nvidia.com/en-us/about-nvidia/company-policies/

Saivasan, R., & Lokhande, M. (2023). Exploring use cases of generative AI and metaverse in financial
analytics: unveiling the synergies of advanced technologies. International Journal of Global
Business and Competitiveness. https://doi.org/10.1007/s42943-023-00082-2
Santa Clara University. (n.d.). Ethics in the metaverse – Markkula Center for Applied Ethics.
https://www.scu.edu/ethics/metaverse/
Saracco, R. (2021, November 12). Metaverse vs Omniverse. IEEE Future Directions.
https://cmte.ieee.org/futuredirections/2021/11/16/metaverse-vs-omniverse/
Telefónica. (n.d.). Welcome to metaverse and web3. Telefónica.
https://metaverso.telefonica.com/en/social-and-ethical-challenges
The ethics of the metaverse. (2022, January 26). Berkman Klein Center.
https://cyber.harvard.edu/story/2022-01/ethics-metaverse
Walton, J. (2023, June 28). Nvidia GeForce RTX 4060 review: Truly mainstream at $299. Tom’s Hardware.
https://www.tomshardware.com/reviews/nvidia-geforce-rtx-4060-review-asus-dual/8
Wikipedia contributors. (2023, May 30). Nvidia omniverse. Wikipedia.
https://en.wikipedia.org/wiki/NVIDIA_Omniverse
Zallio, M., & Clarkson, P. J. (2023). The ethics of the metaverse and digital, virtual and immersive
environments: Metavethics. In Studies in big data (pp. 93–103). https://doi.org/10.1007/978-3-
031-29132-6_6

Our team discussed every single detail to make sure Creatives Planet is the most versatile No coding skills Team discussed every single detail to make sure Creatives Planet is the most versatile No coding skills. Our team discuss every single detail to make sure Creatives Planet is the most versatile No coding skillsAlex jhon martin

Is NVIDIA’s Omniverse ethical? – Professor Edmond C. Prakash