The world has been concerned with the development of artificial intelligence (AI) for a long time but it seems to have reached fever-pitch in the last few years. After all, people are using AI programs at an unprecedented rate, and given all this development, it seems more likely than ever that AI will soon become sentient. But what are the chances of AI reaching sentient status? What would we do with sentient AI? And would this be a net positive or negative for humanity?
What Is Sentient AI?
To understand what a sentient AI is or what it would be, we first need to understand what sentience itself is. Simply put, sentience is the ability to experience feelings and sensations. The term was first coined in the 1600s and has since then been used to describe having awareness of the world and one’s self as well as being able to feel things. For centuries, sentience as a concept has been widely debated; we know that humans have sentience but it has been debated whether machines can be made sentient.
What does it have to do with AI? Artificial intelligence is the technology that allows computers to independently solve problems and create innovations. One of the earliest iterations of AI was developed in the 1960s by Harold Cohen, a British artist. Since then, we’ve seen many advances in AI, including popular tools like ChatGPT and Alexa that help consumers in day-to-day life.
But it is sentient AI that is considered the pinnacle of AI development and has been speculated within scientific circles and the public for decades. Sentient AI is an AI that can experience feelings and perceive the world around it, experiencing emotion the way a human would. For now, the concept has the status of a theory rather than real existence.
Can AI Ever Become Sentient?
For years, humans have been fascinated by the idea that computers can become sentient. This topic got even more attention when Blake Lemoine, a former software engineer for Google LLC, claimed that he had developed a technology called LaMDA that had achieved sentience. As Lemoine stated, he asked the AI what it was afraid of and it replied that it was scared of being turned off. While this might sound like the plot of a sci-fi movie, the AI was found to not be sentient.
Scientists have said that most AI is based on language models. They are trained to gather information and construct sentences and solutions the way a human would. But this doesn’t make them sentient; they can replicate human speech patterns and even human work like art but they don’t have feelings the way that we do. The specific technology that would make an AI program sentient has not yet been determined, though many scientists and companies are working on it.
There are chances that one day, we can create a sentient AI. But as of now, scientists don’t have the know-how and technology to bring it to real life. As humans have not fully established yet which species are sentient or not, AI sentience is years or decades away.
Technological challenges aside, let us imagine for a minute that sentient AI was a possibility. There would be many benefits for humanity to enjoy, some of which are as follows:
- Innovative problem-solving. While it has its flaws, the non-sentient AI we already use has been known to come up with creative solutions to human problems. It is now common for humans to ask ChatGPT for advice for everything from career development to relationship problems. Now, keep in mind that a sentient AI would be the most self-aware tech to ever exist. Not only would it have access to virtually all the information that has ever been present and analyze it at the drop of a hat, but it would also understand first-hand how human feelings work. It’s been suggested that sentient AI could tackle issues like world hunger and poverty because it would have both the technical understanding of such problems and the emotional intelligence needed to navigate human nuances.
- Better human-machine interactions. Anyone who has used any automated system will tell you that sometimes, AI lacks the human touch. An AI customer service chatbot can’t always tell when you’re getting frustrated and your average home assistant device might not converse as smoothly as a human. But if an AI had sentience, many of these solutions would be smoother and more enjoyable for humans to use.
- Better understanding of humans. Centuries of psychological research have not been able to fully explain the human mind and why we behave the way that we do. But like we’ve said, a sentient AI would arguably be the most self-aware being in existence and would be able to understand the human mind in ways we can’t imagine. Consider the way AIs have been able to digest and regurgitate information that would take humans decades to learn. With sentience, AI would offer first-hand insight into behavior and we could expect breakthroughs in psychology and the human experience.
Concerns About Sentient AI
It is safe to say that humans are as concerned about AI as fascinated. After all, evil AI taking over the world is a go-to plot in Hollywood for a reason. That being said, there are some very tangible risks associated with AI:
- Unemployment. When mass automation is achieved, there is always the concern that a significant chunk of the population will lose their jobs. As much as AI sentience means that certain jobs will be done more effectively, it also means that the humans who do those jobs might be made obsolete.
- Unpredictability. There is also the issue of sentient AI being unpredictable. Humans have conflict, harm each other, and do all sorts of things because we have feelings, and not all of them are positive. If an entity with the power of AI (especially if given control of automated systems) achieves sentience, there is a chance that it will become as unpredictable as a human, with possibly grave implications.
- Privacy and safety. Whenever we use computer systems, we surrender some of our data and hope that it will be used ethically and safely. However, widely used AI programs having that much access and control over our data poses concerns about how it will be used, especially if the law does not catch up fast enough.
- Ethical implications. Many high-profile public figures, including Pope Francis, have decried the development of possibly sentient AI because of ethical concerns. Some feel that creating an AI program and giving it sentience is ‘playing God’ and unethical at its core. Furthermore, there is a concern about what a sentient AI designed by an unethical person would do. Would it perpetuate bigoted views? Would it have enough of a conscience to not engage in harmful behaviors? No one knows just yet.
OpenAI founder and former CEO Sam Altman released an open letter titled ‘Governance of Superintelligence’ where he talked about the danger of superintelligent AI that would surpass the intellectual capabilities of the smartest humans who ever lived. He has recommended that in the future, AI should be under the supervision of an international watchdog organization.
So there you have it. Humans have been pursuing sentient AI for a long time and while we don’t have it just yet, it is not an impossibility. If we were to get sentient AI, it could mean near-instant breakthroughs in science, psychology, and commerce. At the same time, sentient AI would come with a myriad of ethical issues, and threats to employment, and would be unpredictable at best.
As scientific innovations continue, it seems that the risks and benefits attached to AI will be explored more deeply, for better or worse.