AI Will Be What We Make of It
What does the future hold with Artificial Intelligence?
Artificial intelligence, or AI, has become one of the most hotly debated and least understood topics in just the last few years. At its most basic level, AI is a technology “that mimics human intelligence to perform tasks and can iteratively improve” itself, according to Oracle, one of the largest computing companies in the world.
Many of us likely had our first exposure to the term AI in the 1991 film Terminator 2: Judgment Day, in which an AI program called Skynet becomes aware of its own existence (also known as the “singularity”) and decides to extinguish all human life. While this fell squarely in the realm of far-future science fiction at the time, hundreds of AI developers signed an open letter to The New York Times in May, warning that AI “could be as deadly as pandemics and nuclear weapons.”
While that may sound extreme, even the faint possibility of such danger is why companies such as Microsoft have made it clear that their priority is “ensuring AI remains under human control” as they explore the technology. That is more critical now than ever because AI research and development have grown explosively in the last few years, bringing the potential to dramatically improve our lives or put us at risk, depending on how it is implemented and who is implementing it.
Perhaps the most famous example of AI right now is ChatGPT, described by its developer as an “AI-powered language model developed by Open-AI, capable of generating human-like text based on context and past conversations,” which the program scrapes from the infinite content on the internet, both good and bad. Introduced just a year ago, it has become the poster child for the capacity of AI to learn, grow and evolve. The program is able to not just mimic many aspects of human conversation, but write essays, perform analyses and write computer code. Similar tools have been developed to generate images, videos and even voices. All of this creates unprecedented opportunities to increase human productivity, theoretically freeing us up to spend less time on mundane tasks and more time on personal pursuits.
Georgia has become one of the frontiers of that AI research and development. Emory and Georgia Tech, for example, recently teamed up to explore how AI can increase the efficacy of diagnoses and treatment for patients. Georgia Tech is also partnering with Southern Regional Technical College (SRTC) to build a technical workforce training incubator and talent pipeline for autonomous and AI technologies.
There is still a colossal amount of work to do to get AI “right,” but it has the potential to revolutionize almost every aspect of human existence.
AI still has a lot to learn, though, and implementing it before it is ready can have serious consequences. Earlier this year, a Black man in Georgia was arrested after a facial recognition program identified him as committing a crime in Louisiana, a state he had never visited.
This is just one of many examples of AI coming up with racist, sexist or plagiarized outputs, which can be especially dangerous when utilized in policing, hiring or educational fields, respectively. There is still a tremendous amount of work to be done to figure out how to train these models appropriately, but right now, guardrails on the technology are minimal to nonexistent, leaving decisions about how and when to create and use it unchecked. Given Congress’ glacial pace of adaptation to new technology (look no further than the ongoing debate around Bitcoin and cryptocurrency), it is unlikely that we will see any kind of national framework on AI any time soon.
Right now, AI is still very much under our control. There is also little risk of it turning into Skynet any time soon. In fact, ChatGPT-4, one of the newest models of the AI bot, went from answering a simple math question right 98% of the time in March 2023 to only 2% of the time in June 2023, according to Fortune. There is still a colossal amount of work to do to get AI “right,” but it has the potential to revolutionize almost every aspect of human existence.
Right now, AI is a little like a child – with the critical difference being that it can absorb almost limitless amounts of information. Much as in a child’s education, the quality and type of information we give it will determine what it looks like when it reaches maturity. We are at an inflection point right now, and it is on us to decide how we want it to look when it’s all grown up.
Tharon Johnson received a Green Eyeshade award for serious magazine commentary for his December 2022 column, “Making Housing Affordable Again.” He can be seen Sunday mornings on The Georgia Gang on Fox 5 Atlanta.
