fbpx
Articles

What a new AI system reveals about our hopes for technology

Open AI’s GPT-3 and remembering to recognize our Creator

/
August 3, 2020

It’s easy for anyone to get caught up in the hype surrounding new technologies. A new innovation often debuts with some helpful benefits and great new features, all of which wow us and lead us to believe that we are on the cusp of something truly revolutionary. Promises are made, and there are countless predictions about what is to come next. But soon after the press conferences fade and the hype dies down, we see these innovations for what they really are—helpful tools with innovative benefits that often do not live up to the hype surrounding their release but also reveal a number of potential misuses, abuses, or failures that we did not account for. Part of this is because we grow accustomed to innovation. But it also happens because we put a level of hope and desire on these technologies to usher in a new era of our world.

OpenAI recently announced their new language model called GPT-3, which is one of the most advanced AI systems in the world to date. This system is truly amazing. It is able to write prose, design and code basic HTML including various mini applications, and even engage in “deep” philosophical conversations about the nature of God and the universe with you. OpenAI released the technical documentation back in May. And according to Morning Brew, GPT-3 “has 175 billion parameters, a 117x increase over its predecessor’s 1.5 billion.” The system was trained on roughly a trillion words. In layman’s terms, it is pretty powerful. The company decided to allow a small group of select users to test out the system, and many shared their experiments online to show off the power of the new system. 

But nearly as soon as people were seeing the immense potential of the system, there came a wave of excitement about what this step forward in AI might mean. “Playing with GPT-3 feels like seeing the future,” a San Francisco-based developer tweeted about this tool. Some even questioned if we were that much closer to human level AI, also known as artificial general intelligence (AGI). However, the dream of a future AGI system is highly debated in computer science circles as well as in philosophy and religion.

Human level intelligence?

This isn’t the first time we have had such utopian dreams with AI. Some of the talk surrounding GPT-3 reminds me of the debut of Google Duplex back in 2018 at their annual developer conference, where Duplex was shown to book a haircut at your local salon or even a table at a restaurant all on its own. There have been countless seasons of grand visions for AI and where we are headed as a society, which ultimately died down over time as we adjusted to our expectations and saw these innovations as encouraging advances but ultimately not as life-altering as promised.

The reality about GPT-3 is that the model is extremely powerful and honestly a good bit of fun based on the users who have been working with it, but this system is no closer to ushering in the famed golden age of AI than any other innovation. This is simply because our current level of AI, known as narrow AI, while powerful and beneficial, is nowhere close to actually understanding the results or products it delivers, nor will algorithmic technology actually be able to achieve the level of general or human-like intelligence. This is because human beings are not machines, even if we often treat each other as mere objects to be manipulated and altered at will. Our minds and consciousness are not simply the result of some chemical reaction or organic algorithm, a view that has been popularized by many thinkers such as Yuval Noah Harari or Ray Kurzweil.

The depth of the human experience

This point has been highlighted by many prominent thought leaders over the years, such as renowned mathematics professor John Lennox of Oxford, the late philosopher Roger Scruton, and many computer scientists like Rosalind Pichard and Joanna Ng. This summer, I spent some time digging into Scruton’s works On Human Nature and The Soul of the World, in which he shows how a naturalistic understanding of the world fails to account for the depth of humanity in terms of our conscious experiences, emotions, moral agency, and even how we see each other as unique beings in this world. This reductionistic view of humanity is often behind the pursuit of the famed rise of AGI, because if there is nothing unique about humanity, then we should be able to recreate human intelligence and experience in a digital form.

Scruton describes one aspect of the uniqueness of humanity as the presence of subjective experiences, or the I/You paradigm, as one of the main differences between how subjects (like you and I) operate in a world of objects (like that of technology). This is one of the reasons we question the nature of ethics, our identity, and even the presence of God himself. Even the animal kingdom doesn’t experience the world as humanity does. We were created unique by God himself as his image bearers (Gen. 1:12-28).

There is a common misconception that our personhood can be derived simply from the material, which leads humanity down dangerous paths of believing that we are less valuable than we really are and overvaluing technology as if it somehow has the potential to become our equal or even surpass us in terms of utility or dignity.

Many may have missed how quickly people acknowledged that this GPT-3 model did not actually exhibit any of the signs of actual human level intelligence, even if the system could do things that were previously unbelievable for an AI system to do. But this longing for a system to create AGI reveals something a bit ironic about our longings and desires as humans that we shouldn’t miss. We often seek to humanize our creations, i.e., technologies, all the while dehumanizing ourselves. With our desire to be like God and create something in our image, we end up having to dumb ourselves down and treat ourselves as if we are merely machines rather than uniquely created image-bearers of the living God.

While many will continue to claim that faith and science are simply at odds with one another and that AGI development is just around the corner, Christians can remember and have hope that even with our wildest attempts or innovations, we are simply not able to change our own human nature nor create something like ourselves. As amazing new technological innovations continue to rise, we can step back and praise God for the incredible, talented people creating these tools rather than focus on some desire to create something on par or even greater than ourselves.

These innovations can be used for immense good, but we also must remember that they will be misused and possibly even become objects that we put our hopes in instead of God himself. We may trick ourselves into believing that it is possible to reach AGI or even create an AI system that can pass the famed Turing test, but we simply are not able to define, alter, or manipulate our humanity and personhood to feed these longings. We are God’s creatures and must never forget how we are called to live in this world—always recognizing our creatureliness and fixing our gaze on the Creator of all life and everything in the cosmos.

Jason Thacker

Jason Thacker serves as chair of research in technology ethics and creative director at ERLC. In his role as creative director, he oversees the communications team, including all creative design projects.  His book, The Age of AI: Artificial Intelligence and the Future of Humanity, released March 2020 with Zondervan. He is a graduate … Read More