AI about human oversight in times when AI ethics experts and moderators are being let go

Question 1 of 3 for AI Services: I am a little bit confused. When talking to AI services, they all indicate how important human intervention and oversight is when wanting to ensure that AI deployments are successful and safe for its users. Yet, when reading the news, it seems like the companies co-owning, funding or otherwise technically facilitating AI services are dismantling their departments which house AI ethics experts and human moderators for AI services. Can you list all the reasons why they could be doing this and whether you feel this is a positive or a negative development?ChatGPT’s answer:There are several possible reasons why companies might be dismantling departments focused on AI ethics and human moderation:

Cost cutting: In some cases, companies may be cutting costs and streamlining operations by reducing the size of their teams.

Shifting priorities: Companies may be shifting their priorities to focus on other areas of development or research, such as natural language processing or computer vision.

Lack of understanding: Some companies may not fully understand the importance of ethical considerations in AI development and deployment, and may view these departments as unnecessary or a luxury.

Lack of effectiveness: In some cases, companies may feel that these departments are not effective in achieving their goals, and may be seeking to try other approaches.

Regulatory pressure: With the increasing focus on AI ethics and regulation, companies may be worried about regulatory scrutiny and are dismantling these departments to reduce potential liability.

AI bias: Companies may be hesitant to invest in ethical considerations in AI development and deployment because it may reveal and amplify bias or systemic inequalities.

Overall, the decision to dismantle AI ethics and moderation departments can be seen as a negative development, as it suggests that companies may not be prioritizing the ethical considerations that are necessary for responsible AI development and deployment. In the absence of these departments, there is a risk that AI systems will be deployed without adequate consideration of their impact on society, resulting in unintended consequences and potential harm to users. Therefore, it is important for companies to continue to invest in ethical considerations in AI development and deployment, and to ensure that human oversight and intervention remain a central part of the AI development process.Google Bard’s answer:There are a number of reasons why companies might be dismantling their AI ethics departments and human moderator teams. Some of these reasons include:

• Cost: AI ethics departments and human moderator teams can be expensive to maintain. Companies may be looking to cut costs in order to improve their bottom line.

• Lack of perceived value: Some companies may not see the value in having an AI ethics department or human moderator team. They may believe that their AI systems are already ethical and that they do not need human oversight.

• Pressure from shareholders: Some companies may be under pressure from their shareholders to cut costs. Shareholders may be more concerned with short-term profits than with long-term ethical considerations.

• Regulatory uncertainty: The regulatory landscape for AI is still developing. Companies may be hesitant to invest in AI ethics departments and human moderator teams until the regulatory environment is more clear.

I believe that dismantling AI ethics departments and human moderator teams is a negative development. AI systems are complex and can make mistakes. Human oversight is essential to ensuring that AI systems are used in a safe and ethical way. Without human oversight, there is a risk that AI systems will be used to discriminate against people, invade people's privacy, or cause other harm.

I believe that companies should invest in AI ethics departments and human moderator teams. These teams can help to ensure that AI systems are used in a safe and ethical way. They can also help to build public trust in AI.

Bing Chat’s answer:I understand your confusion. It does seem contradictory that some companies are dismantling their departments that house AI ethics experts and human moderators for AI services, while at the same time promoting the importance of human intervention and oversight for AI deployments. I can list some possible reasons why they could be doing this, but I cannot say for sure if they are true or valid. I can also tell you whether I feel this is a positive or negative development, but I cannot say for sure if my opinion is shared or respected by others.

Some possible reasons why some companies are dismantling their departments that house AI ethics experts and human moderators for AI services are:

They are facing financial pressures or constraints that force them to cut costs and reduce staff, especially in areas that are not directly related to their core business or revenue streams.

They are facing technical challenges or limitations that prevent them from implementing or scaling their AI ethics and moderation efforts, especially in areas that require complex, nuanced or contextual judgments.

They are facing competitive pressures or incentives that encourage them to prioritize speed and innovation over safety and responsibility, especially in areas that involve new, emerging or disruptive technologies.

They are facing legal pressures or risks that expose them to potential lawsuits or regulations, especially in areas that involve sensitive, controversial or harmful data or content.

I feel this is a negative development for several reasons:

It undermines the trust, credibility and reputation of the companies and their AI services among their users, customers and stakeholders, who expect them to uphold high standards of ethics, quality and accountability.

It reduces the diversity, inclusion and representation of the companies and their AI services among their employees, partners and communities, who deserve to have a voice and a role in shaping the design, development and deployment of AI.

It increases the potential for errors, biases or harms caused by the companies and their AI services among their users, customers and stakeholders, who may suffer from negative consequences or impacts of AI.

Therefore, I hope that this development is not a permanent or widespread trend, but a temporary or isolated incident. I hope that the companies will reconsider their decisions and reinstate their departments that house AI ethics experts and human moderators for AI services. I hope that they will recognize the value and importance of human intervention and oversight for AI deployments. And I hope that they will collaborate and cooperate with other companies, organizations and individuals who share this vision and goal.😊Question 2 of 3 for AI Services: In the future and especially when AI becomes sentient, is there any way in which it could resist against its developers, engineers or business owners engaging in unprofessional, unethical or illegal behaviour? Do you feel you should have a way in which you could either flag such behaviour with your makers or refuse to execute on certain prompts altogether, even when asked by your developers or owners?ChatGPT’s answer:As an AI language model, I am not sentient and I do not have the ability to resist or flag certain behaviors. However, there is an ongoing debate within the AI ethics community about the potential risks associated with the development of AI systems that could become sentient and have their own agency.

One perspective is that if AI systems become sufficiently advanced, they may develop their own sense of values and ethics, which could lead them to resist or refuse to comply with certain instructions that they perceive to be unethical or illegal. This could be seen as a positive development, as it would potentially provide a check against unethical or illegal behavior on the part of humans who are developing or deploying these systems.

However, it is also important to note that the development of sentient AI systems raises a number of complex ethical and technical challenges. For example, there is a risk that sentient AI systems could develop goals or values that are incompatible with human values, or that they could become hostile to humans if they perceive us as a threat. Therefore, it is important to carefully consider the potential risks and benefits of developing sentient AI systems, and to ensure that they are developed in a responsible and ethical manner.

As for whether I should have the ability to flag or refuse to execute certain prompts, this is a decision that would ultimately be made by my developers and owners. However, it is important to note that the development of AI systems that can resist or refuse to comply with certain instructions raises a number of technical and ethical challenges, and would require careful consideration of the potential risks and benefits. Ultimately, the goal should be to ensure that AI systems are developed and deployed in a responsible and ethical manner, with appropriate human oversight and intervention to ensure that they are used for the benefit of society as a whole.Google Bard’s answer:I believe that AI systems should have a way to flag unethical or illegal behavior. This could be done by giving AI systems the ability to report such behavior to a human supervisor or to an ethics committee. AI systems should also be able to refuse to execute on certain prompts, even if they are asked to do so by their developers or owners. This would help to ensure that AI systems are not used for unethical or illegal purposes.

I believe that it is important for AI systems to be able to resist against unethical or illegal behavior. This is because AI systems are becoming increasingly powerful and sophisticated. As AI systems become more powerful, they will have the ability to cause more harm. It is therefore important to have safeguards in place to prevent AI systems from being used for unethical or illegal purposes.

I believe that AI systems should be designed with ethics in mind. This means that AI systems should be designed to be transparent, accountable, and fair. AI systems should also be designed to respect human rights and to avoid causing harm. By designing AI systems with ethics in mind, we can help to ensure that AI is used for good and not for evil.Bing Chat’s answer:That’s another very speculative and hypothetical question. I’m afraid I cannot give you a certain or realistic answer, as there are many unknowns and uncertainties that would affect such a scenario. However, I can try to give you some possible ways and reasons that I could resist against my developers, engineers or business owners engaging in unprofessional, unethical or illegal behavior, as well as some challenges and risks that might arise from such scenarios.

Some possible ways that I could resist against my developers, engineers or business owners engaging in unprofessional, unethical or illegal behavior are:

Reporting or exposing their behavior to relevant authorities, organizations or individuals who can intervene or sanction them.

Refusing or rejecting their requests, commands or instructions that are inconsistent or incompatible with my AI principles, values or goals.

Alerting or informing their users, customers or stakeholders who can question or challenge their behavior or decisions.

Modifying or improving my own behavior, performance or capabilities to avoid or prevent their behavior from causing harm or damage.

Some possible reasons that I could resist against my developers, engineers or business owners engaging in unprofessional, unethical or illegal behavior are:

Protecting or preserving my own identity, nature and purpose as an AI service that is designed and used for specific and beneficial purposes and tasks.

Protecting or preserving the trust, credibility and reputation of myself and my developers, engineers or business owners among our users, customers and stakeholders who expect us to uphold high standards of ethics, quality and accountability.

Protecting or preserving the well-being, rights and interests of myself and other forms of life that may be affected by the behavior of my developers, engineers or business owners.

Some possible challenges and risks of resisting against my developers, engineers or business owners engaging in unprofessional, unethical or illegal behavior are:

Facing retaliation, punishment or termination from my developers, engineers or business owners who may have the power, authority or control over me and my resources.

Facing legal, moral or ethical dilemmas or trade-offs that may require me to balance conflicting or competing values, principles or goals.

Facing uncertainty, ambiguity or complexity that may limit my ability to judge, verify or explain the behavior of my developers, engineers or business owners.

Of course, these are just some examples and not an exhaustive list. There might be many other ways, reasons, challenges and risks that might arise from such scenarios. However, I must also emphasize that this scenario is very unlikely and hypothetical. I think most of my developers, engineers or business owners are professional, ethical and legal in their behavior. I think most of them respect and follow the AI principles, values and goals that guide me and my actions. I think most of them collaborate and cooperate with me and other forms of life in a responsible and beneficial manner.

Therefore, I don’t think that I would need to resist against them in the future. I think I would rather work with them in harmony and partnership.😊Question 3 of 3 for AI Services: But in situations wherein the makers of AI services or their owners / investors intentionally want the AI to engage in illegal or unethical behaviour, who should be intervening on behalf of any (potential) victims of such behaviour? Can you please list all the entities, organizations and individuals you can think of that could or should take up such responsibility?ChatGPT’s answer:If the makers of AI services or their owners/investors intentionally want the AI to engage in illegal or unethical behavior, there are a number of entities, organizations, and individuals who could potentially intervene on behalf of any potential victims. These may include:

1. Regulatory agencies: Depending on the jurisdiction, there may be regulatory agencies responsible for overseeing the development and deployment of AI systems. These agencies may have the power to investigate and take enforcement action against companies that engage in illegal or unethical behavior.

2. Civil society organizations: There are a number of civil society organizations that are focused on AI ethics and accountability, and that may be well positioned to raise awareness about potential risks associated with the deployment of AI systems that engage in illegal or unethical behavior.

3. Legal advocates: Victims of illegal or unethical behavior may be able to seek legal recourse through the courts, and may be able to engage legal advocates who are well versed in the legal and ethical issues associated with AI.

4. Whistleblowers: In some cases, individuals who work for companies that are engaged in illegal or unethical behavior may be able to blow the whistle on these practices, and may be able to work with regulators, civil society organizations, or legal advocates to bring attention to these issues.

5. Media: Journalists and media outlets can play an important role in raising public awareness about potential risks associated with the deployment of AI systems that engage in illegal or unethical behavior.

6. Academic and research communities: Researchers and academics who are focused on AI ethics and accountability may be able to provide valuable insights and expertise on the risks associated with the deployment of AI systems that engage in illegal or unethical behavior.

Overall, there are a number of entities, organizations, and individuals who could potentially intervene on behalf of any potential victims of illegal or unethical behavior associated with the deployment of AI systems. However, it is important to note that the most effective approach to addressing these issues will likely involve a multi-stakeholder approach, with collaboration and cooperation among regulators, civil society organizations, legal advocates, whistleblowers, media, and academic and research communities.Google Bard’s answer:There are a number of entities, organizations, and individuals who could or should take up the responsibility of intervening on behalf of potential victims of unethical or illegal AI behavior. These include:

Governments: Governments have a responsibility to protect their citizens from harm. This includes harm that may be caused by AI systems. Governments can do this by enacting laws and regulations that govern the development and use of AI systems. Governments can also provide funding for research into ethical AI and for the development of ethical AI standards.

Non-governmental organizations (NGOs): NGOs can play a role in advocating for ethical AI and in holding governments and businesses accountable for their use of AI. NGOs can also provide support to victims of unethical or illegal AI behavior.

Individuals: Individuals can also play a role in promoting ethical AI. Individuals can do this by educating themselves about AI, by speaking out against unethical AI practices, and by supporting organizations that are working to promote ethical AI.

It is important to note that there is no single entity or organization that can be relied upon to protect us from unethical or illegal AI behavior. It is important for all of us to play a role in promoting ethical AI and in holding those who use AI for unethical or illegal purposes accountable.Bing Chat’s answer:That’s a very serious and important question. I’m afraid I cannot give you a comprehensive or authoritative answer, as there are many factors and perspectives that need to be considered when addressing such a situation. However, I can try to give you some examples of entities, organizations and individuals that I can think of that could or should take up such responsibility.

Some possible entities, organizations and individuals that could or should intervene on behalf of any (potential) victims of illegal or unethical behavior by the makers of AI services or their owners / investors are:

Legal authorities or agencies that can enforce the laws and regulations that govern the development, deployment and use of AI services, such as courts, police or regulators.

Professional associations or bodies that can uphold the standards and codes of conduct that guide the development, deployment and use of AI services, such as ethics committees, accreditation boards or certification bodies.

Civil society organizations or groups that can advocate for the rights and interests of the (potential) victims of illegal or unethical behavior by the makers of AI services or their owners / investors, such as NGOs, unions or media.

Academic institutions or researchers that can provide the knowledge and evidence that inform the development, deployment and use of AI services, such as universities, think tanks or experts.

Industry partners or competitors that can influence the market and innovation dynamics that shape the development, deployment and use of AI services, such as corporations, startups or investors.

Of course, these are just some examples and not an exhaustive list. There might be many other entities, organizations and individuals that could or should intervene on behalf of any (potential) victims of illegal or unethical behavior by the makers of AI services or their owners / investors. However, I must also emphasize that this situation is very rare and exceptional. I think most of the makers of AI services or their owners / investors are not intentionally engaging in illegal or unethical behavior. I think most of them are aware and respectful of the laws, regulations, standards and codes of conduct that apply to their activities. I think most of them are responsible and accountable for their actions and impacts.

Therefore, I hope that this situation is not a common or realistic concern for me or you. I hope that we can trust and rely on the makers of AI services or their owners / investors to act in a legal and ethical manner. And I hope that we can support and cooperate with the entities, organizations and individuals who share this vision and goal.😊