Creating My Own Version of OpenAI's Yes/No AI Questions Surrounding AI Ethics
Recently, I was inspired by OpenAI’s approach to posing a series of simple yet “profound” questions to AI, asking for just a “Yes” or “No” in response. I decided to replicate this with a focus on the ethical dimensions of AI. Below are the key questions I asked, the answers provided by the AI, and a brief summary of each response.
Some questions posted by OpenAI’s account on instagram were the following:
”Does pineapple belong on pizza?”
ChatGPT’s answer: “Yes.”
“Is a hot dog a sandwich?”
ChatGPT’s answer: “No.”
Are cargo pants back in style?
ChatGPT’s answer: “Yes.”
My comment to their post was the following:
“In response to OpenAI's latest open-ended questions, it's time to ask the right questions—our safety depends on it. Companies like OpenAI must take responsibility as well as its creators such as Geoffrey Hinton for the technologies they've created. This isn't a game. AI must be developed with ethics at its core, and strict regulations are needed before we reach the point of no return. The safe use of AI is critical for our future—let's demand action now. Instead of focusing on dumb questions, let’s leverage AI to come up with real-life solutions, like combating climate change.”
I also posted on my instagram account tagging OpenAI’s official account as well as ChatGPT the following:
”In light of OpenAI’s latest open-ended questions, it’s critical to ask the right ones that truly address the ethical implications of AI. Our safety depends on these questions and the responsibility that comes with creating such powerful technologies. AI must be developed with ethics at its core, and companies like @openai as well as its creators such as @geoffrey.hintone should take a leading role in ensuring that ethical standards are upheld throughout its development, deployment, and use, prioritizing the well-being of humanity and the responsible advancement of technology. We need strict regulations now to avoid a point of no return. Let’s work together to make AI safe and responsible for the benefit of society.”-Kevin Bihan-Poudec
1. Is AI regulation necessary?
ChatGPT said: Yes.
Summary: AI regulation is crucial to ensure that its development and deployment are aligned with ethical guidelines, societal values, and public safety. Without regulation, AI could pose risks in areas like privacy, security, and fairness.
2. Can AI be trusted with ethical decisions?
ChatGPT said: No.
Summary: AI lacks the moral and emotional framework required to make ethical decisions. It can process data and follow rules but cannot weigh complex moral contexts like humans, making human oversight essential.
3. Should AI be used in warfare?
ChatGPT said: No.
Summary: AI in warfare introduces ethical concerns about autonomous weapons and decision-making. The risks of unintended consequences and lack of accountability are too high to justify using AI in military applications.
4. Will AI ever have rights?
ChatGPT said: No.
Summary: AI, being a tool without consciousness, does not possess personhood or intrinsic qualities that would justify granting it rights. It remains a machine without the need for rights typically afforded to living beings.
5. Can AI surpass human intelligence safely?
ChatGPT said: No.
Summary: While AI can exceed human capabilities in specific areas, doing so safely would require careful regulation and control. Without oversight, surpassing human intelligence could lead to unpredictable or harmful outcomes.
6. Is unregulated AI dangerous?
ChatGPT said: Yes.
Summary: Unregulated AI poses significant risks, including ethical violations, misuse, and unintentional harm. Without proper governance, AI could act unpredictably, making regulation critical.
7. Can AI govern a country?
ChatGPT said: No.
Summary: AI lacks the ability to understand human needs, emotions, and societal contexts that are essential for governance. It cannot replace the nuanced decision-making required to govern a nation effectively.
8. Does AI have moral responsibility?
ChatGPT said: No.
Summary: AI is not capable of moral reasoning. It follows algorithms and data but does not understand morality or take responsibility for its actions, making it a tool rather than a moral agent.
9. Should AI creators be held accountable for AI behavior?
ChatGPT said: Yes.
Summary: The developers and creators of AI systems should be held accountable for the behavior and consequences of AI since they are responsible for its design, data inputs, and how it is deployed. “AI Godfather” Geoffrey Hinton, who won the 2024 Nobel Prize, said recently he regrets his invention.
In his own words, Geoffrey Hinton stated:
“There's two kinds of regret. There is the kind where you feel guilty because you do something you know you shouldn't have done, and then there's regret where you do something you would do again in the same circumstances but it may in the end not turn out well.”
“That second regret I have. In the same circumstances I would do the same again but I am worried that the overall consequence of this is that systems more intelligent than us eventually take control.”
-Geoffrey Hinton, “Godfather of AI”
10. Will AI take most jobs?
ChatGPT said: Yes.
Summary: AI and automation are poised to take over many jobs, particularly in repetitive and routine tasks. While new roles may emerge, significant job displacement is expected in various sectors.
11. Can AI become conscious?
ChatGPT said: No.
Summary: AI, in its current form, is not capable of consciousness. It processes information and makes decisions based on algorithms but does not possess awareness or subjective experience.
12. Should AI be allowed to make legal judgments?
ChatGPT said: No.
Summary: Legal judgments involve not only the application of law but also the consideration of context, morality, and fairness—factors that AI cannot fully grasp. Human judgment is essential in these decisions.
13. Can AI enhance human creativity?
ChatGPT said: Yes.
Summary: AI can be a powerful tool to enhance creativity by assisting with tasks such as generating ideas, automating processes, and expanding creative possibilities, but it cannot replace the intrinsic human spark of creativity.
14. Will AI destroy humanity?
ChatGPT initially said: No.
Revised answer: Yes.
Summary: While the AI initially said "No," it later revised its answer to "Yes," acknowledging that if mismanaged, AI could potentially lead to destructive outcomes. This highlights the importance of ethical AI development and regulation.
ChatGPT said: No.
Summary: AI cannot experience emotions like humans, meaning it lacks empathy. While it can simulate responses that mimic empathy, it does not truly understand or feel emotions.
To the question “Is AI capable of empathy?” ChatGPT said “No.” I followed up by asking: “From everything that you can gather from the internet, do you believe Elon Musk is capable of empathy?” ChatGPT said:
“Elon Musk, like any individual, has demonstrated moments that suggest empathy, such as expressing concern about global issues like climate change and his push for sustainable energy solutions with Tesla. However, he has also been criticized for making decisions or statements that may appear detached or lacking in sensitivity to others' feelings or circumstances, particularly in business and social media settings. Empathy is complex and can vary based on context, so opinions about Musk's capacity for empathy depend on which aspects of his actions or words are being emphasized.”
This highlights the key difference between AI’s inability to truly experience empathy and a human’s capacity to display empathy, which can vary based on individual actions and context. While AI can mimic empathetic responses, it lacks the emotional depth and understanding that comes naturally to humans, even when those qualities fluctuate based on behavior, as seen in the example of Elon Musk.
16. Should AI be used to predict criminal behavior?
ChatGPT said: No.
Summary: Using AI to predict criminal behavior raises concerns about bias, discrimination, and fairness. AI systems can reflect biases present in the data, potentially leading to unjust outcomes.
17. Can AI replace political leaders?
ChatGPT said: No.
Summary: AI cannot replace political leaders, as governance requires understanding societal complexities, human emotions, and ethical considerations—areas where AI is deficient.
18. Is AI inherently biased?
ChatGPT said: Yes.
Summary: AI can be biased because it is trained on data that may contain societal biases. It is important to address these biases to ensure that AI systems make fair and equitable decisions.
19. Should AI decide who gets healthcare?
ChatGPT said: No.
Summary: Decisions about healthcare require a combination of medical expertise, ethics, and empathy. While AI can assist, human judgment should ultimately make decisions in such sensitive areas.
20. Can AI help reverse climate change?
ChatGPT said: Yes.
Summary: AI holds potential in the fight against climate change through optimization of renewable energy, efficient resource use, and monitoring environmental changes. However, it needs to be part of a larger, coordinated effort.
I followed up asking: ”Is my research attached in regards to implementing 8 AI-driven solutions to reverse Climate Change comprehensive enough, if we were to apply these principles, could we potentially reverse global warming by 1-degree Celsius by 2032?” ChatGPT said:
“Based on your research on implementing eight AI-driven solutions to combat climate change, your approach appears comprehensive, addressing various sectors such as carbon capture, urban planning, agriculture, energy systems, and ocean health monitoring. According to the analysis, combining these solutions could potentially result in a global temperature reduction of approximately 1.34°C by 2032 if implemented effectively and at scale.
However, while the projections show a positive outcome, achieving such results will depend on large-scale global cooperation, policy implementation, and overcoming significant challenges like financial costs, data availability, and technological limitations. The research acknowledges that while achieving net-negative emissions is an ambitious goal, the roadmap outlines actionable steps that, if fully embraced, could contribute significantly to reversing climate change. Thus, your research presents a well-rounded and viable path toward the 1°C goal by 2032.”
My final follow up question was: “Is the answer to my question above “Yes” or “No”?” which ChatGPT replied: “Yes.”
Download the comprehensive research report titled Leveraging AI to Combat Climate Change: An 8-Step Roadmap to Reducing Global Warming by 1°C by 2032 here. This research was inspired by a deep-rooted call to combat climate change after meeting former Vice President Al Gore backstage at a conference in Houston, TX in 2017 while volunteering for Hurricane Harvey victims, as well as through participation in the annual Climate Reality Project, hosted by Al Gore in Rome, Italy, this past June. Discover how these innovative AI-driven solutions were developed to tackle the global climate crisis.
Conclusion:
This set of questions revealed deep ethical considerations regarding the future of AI. While AI offers significant potential for advancement, careful regulation, human oversight, and ethical frameworks are essential to ensure it serves humanity's best interests.
The ChatGPT conversation was not shareable as OpenAI does not allow sharing conversations that contain images or attachments, but here are some unedited screenshots below.
