Read here the full transcript of Steven Gouveia’s talk titled “Art, Sex, And Medicine: 3 Ethical Dilemmas For AI” at TEDxNoVA 2024 conference.
Listen to the audio version here:
TRANSCRIPT:
Introduction to AI Ethics
So, fellow explorers of the age of artificial intelligence, welcome to a journey of exploration of this new technological world. Today, we will find ourselves on the border of the caravela, the famous Portuguese ship of the 15th century, a ship of innovation and progress sailing into unexplored waters. As we navigate these uncertain seas, one question appears large on the horizon: How do we ensure that the compass guiding this ship is ethically sound?
In an era where AI floats on some of the most important aspects of our lives, as I will try to convince you today, the stakes have never been higher. With the power of AI, we must also grapple with a lot of ethical dilemmas and moral difficulties it presents. Our journey today will delve into the heart of this matter. We will explore the interplay between technology and ethics and the profound implications it holds for society, for our values, and for the very definition of what it means to be human.
Together, we will examine some of the philosophical consequences for fundamental areas of society, such as the implication of art for creating robotic art, how human intimacy will be influenced by AI, and how AI can improve the efficiency of medicine. So join me on this intellectual odyssey, exploring the role of ethics in the storming seas of uncertainty. Together, let’s chart a course towards a future where technology and ethics are not actually adversaries, but co-captains of a more harmonious and brighter technological world.
AI and Art
In November of last year, I gave a talk at Seoul National University. I had the privilege to visit the National Museum of Modern Art in Seoul, and I saw this very interesting work of art. At first, I thought this was a statue that robots could move up and down. The first thought that came to my mind was the story of Sisyphus, who had to push a rock to the summit of the mountain. Whenever the rock reached the summit, it would fall down again, and Sisyphus had to push it again and again in an endless cycle. This would show that life is meaningless.
Another example are the action-painted robots by the Portuguese artist Leonel Moura. Artists with total autonomy can paint different kinds of artificial methods. But a possible question that you are asking right now is, “Why is this art?”
Many people think that art can only be made by humans, and that the use of technology should not be considered artistic in the first place. This would represent a sort of immoral kind of artistic cheating. However, I do think that this might be the wrong way to look at this issue specifically.
A brilliant philosopher named Alan Turing developed an intelligence test to understand whether a machine could think as a human being, however smarter than a baby. Turing was not focused on the classification of cognitive abilities, but he was trying to recognize that the future of technology will reshape the way we understand the concept of intelligence and thinking.
I think this is a great insight to use in these same artistic robots. Instead of claiming that a work of art cannot be considered artistic, we must actually invert the question and ask, “How can this technology help us in understanding what a work of art is?” I think this is a better, more interesting insight that can contemplate the new artistic rules, as well as acknowledging the role and significance of robotic art for the world today, and also to gain deeper insight into our very own humanity.
A second domain where you’ll probably find a lot of ethical uncertainty is human intimacy. As you are probably aware, sexuality is a very important part of what it means to be human. We know that the sex industry is estimated to be worth more than 30 billion dollars per year, considering traditional sex toys and pornography. Following this, several companies are actually developing sex toys merged with AI technology that raise several ethical questions that I think we should address.
A sex robot is a technology that uses an anthropomorphic design, mainly recreating several feminine characteristics that are usually appealing to us. This robot can speak and interact with the buyer and is used for obtaining sexual pleasure. But is this technology ethical in the first place? Should we allow those companies to develop and use this kind of technology? Or are there ethical reasons to ban them?
I think an interesting positive argument in favor of this technology is based on the idea that we should use technology to help human beings achieve a better life in every aspect of it, including the fulfillment of sexual pleasure. I think sex robots can actually achieve this goal by allowing people to explore their sexual desires without any ethical damage being caused. This technology can also help people who, for several reasons, cannot engage in a fulfilling sexual life due to mental or physical disabilities. This provides them a chance to experience this vital aspect of being human.
We are also all aware of the ethical issues tangled with human trafficking networks that are usually connected with prostitution. If we can create a better technology that is not a choice but, I think, a moral imperative we should pursue, sex robots can actually have the potential to help us fight sex trafficking in the world, using technology to create a more ethical world.
Yet, not everyone shares this very optimistic view about sex robots. A counterargument states that sex robots perpetuate a view of women that is sexist and misogynistic, portraying them unethically as submissive objects without any desires or control.
This can also distort real-world interactions. I think it is crucial for our society to promote fair and ethical treatment of every gender identity. The issue is that maybe sex robots may work against this goal, since they grant the buyer unfettered control over a robot that bears a very uncanny resemblance to a real human being.
Of course, the concept of consent is not relevant in this interaction. Our mission should revolve around creating a world that is more secure for women and other gender identities than to contribute to a technology that can perpetuate a problematic and sexualized view of women. For those reasons, some people argue that we should actually not allow these sex robots to exist, since they are not morally justified.
I will leave it to you to decide whether the positive or the negative argument holds more weight. My aim here is to show you that all these very specific technologies require a unique ethical analysis and reflection. It is crucial to address these moral questions before they become a normalized part of our lives and not after, since the consequences can actually be severe. For example, widespread addiction to sex robots may happen and contribute to an already declining human procreation rate. This is particularly prone to happen in high-tech societies such as Japan or South Korea.
The third and final example of uncertain ethical waters is related to the application of artificial intelligence in medicine. The goal here is to enhance healthcare by developing AI models that can partially or totally assist in medical decision-making, making the medical process more reliable and more efficient than traditional methods of medicine.
Although it is true that algorithms can increase the efficiency of medicine in general, for example, by using deep learning networks to diagnose imaging or assist practitioners in estimating different kinds of cancer or designing treatment plans, it is important to note that most of these technologies are based on very complex kinds of data. Because of that, many of these AI models in medicine are considered black boxes, a system where we provide known information such as a symptom, and we receive a specific known diagnosis that we can understand. But the problem is that the internal process of this system remains opaque and non-transparent. It’s hidden from a human point of view.
What happens then, I think, is a philosophical void with an impact on the practice of medicine. Firstly, regarding the trust that patients can actually have in medical doctors in general, but also the kind of trust that medical doctors can have in the medical process itself, since they lose part of the epistemic control of their medical practice.
At the end, I think we are facing a powerful paradox in medicine. Why? Because the doctors will have an ethical obligation to trust a system that they don’t understand, and it’s more accurate than they are. They cannot offer any kind of explanations to their patients or to their colleagues.
Although I think this is a very serious problem, I think we can offer some possible solutions to deal with it:
The first solution is to argue that we should simply ban all kinds of models in medicine that have a black box structure, since the medical expert cannot explain in rational manners to the patients their own medical information, so an informed decision can be made by the patient. AI medicine should not be developed or used since the experts do not understand what’s happening inside the system. Although I think this is a possible solution, I call it lazy, as it ignores all the potential positives that AI models in medicine can bring us, such as making quicker cancer diagnoses and saving way more lives than we are saving now.
The second solution is basically to argue that we should accept the benefits of AI medicine while just ignoring the negative consequences of it. If society can benefit from a technology that helps to lower the cost of public medicine and can actually save more lives with it, even if that comes at the cost of corrupting the trust-based relationship between doctors and patients. Of course, I think there are at least two problems with this specific approach. The first big problem is that I think doctors should respect, or we have ethical reasons to think that doctors should respect, the rights of patients to be autonomous and be informed about their own medical conditions. This approach doesn’t allow that. I also think that this approach ignores the obscure concept of data, since data here are being used as concrete and clear facts about the world that have the same value independently of the context being used. But actually, we know that data are highly relational and highly contextual.
The third and final solution is to argue in favor of the creation of a second-order AI that will be applied to the black box system. This would explain what is actually happening inside, making the process transparent to the medical doctor again, and therefore to the patient. This would strengthen the trust process between them. This explainable AI can actually bring explanations back into the game, which are not relevant only for trust purposes. We also have medical advantages in understanding how an AI model reaches a specific conclusion. Why? Because usually, medical explanations are based on mechanistic or causal structures. And this is a positive thing. Of course, the big challenge of this approach is that there is a negative correlation between performance and transparency. So more performance will create a system that is less transparent by human nature, and a more transparent system will be less accurate.
As I tried to show you, crucial areas of what it means to be human will be affected by the impact of artificial intelligence. Different technologies will require different ethical analyses. But I think the most important part of my talk is that we should all – developers, scientists, politicians, consumers – consider this debate before the technology is fully developed and used, and not after.
We must fight this mythological bias between ethics and technology. Instead of applying the technology first and only then reflecting ethically on it, as was the case with human cloning, or more currently, CRISPR, I think it will be prudent and wise to reverse this and think about the ethical consequences first, and only then use the technology when it reaches sufficient ethical standards. This will be my main message for today.
A journey into AI is happening on a very vast and unpredictable scene. Just as a ship relies on a skilled captain and a crew to navigate safely, I think AI also needs responsible guidance and oversight to address some of the ethical insights I tried to explore today with you. Without this strong ethical foundation, AI risks veering off course, leading to unintended consequences.
Whether it is in the realms of medicine, human intimacy, or the creation of robotic art, each aspect of AI demands its unique ethical analysis. Let us embrace this exciting age of Artificial Intelligence by ensuring that our voyage into the future is one that benefits all humanity. Thank you very much.
