This past year, we’ve seen several new, flashy pieces of hardware and software come out, namely Windows 11, iOS 15, iPhone 13, Pixel 6, and the Samsung Flip3. But most importantly, what came out of 2021 were new ideas of what technology’s role in society should be. As personal data grows in value, artificial intelligence (AI) improves significantly at pretending to be human, and gene editing becomes a part of reality, technology is no longer a question of whether we could but rather whether we should. Technology is a powerful tool, and it is up to us to decide how to harness its potentially devastating capabilities.

Without further ado, here is a list of the most promising applications of technology that came out of 2021 and the CES 2022 conference.

mRNA Vaccines

Since the start of the COVID-19 pandemic, mRNA, or messenger RNA (ribonucleic acid) has been given a lot of attention in the scientific community—and for good reason. mRNA allows scientists to essentially program cells to make proteins using an arbitrary mRNA sequence, as this RNA is read by the ribosomes and used to synthesize proteins. Viruses can use this to hijack the cell system, injecting malicious mRNA into a cell and using the resulting proteins to propagate the spread of the virus. Compared to previous vaccines, which use weakened versions of disease-causing pathogens, mRNA vaccines “tell” cells to make a portion of the virus (this will be a specific protein or antigen; in the case of the SARS-CoV-2 virus, this is the spike protein); upon seeing the protein or antigen, the immune system begins producing antibodies and learns to recognize that pattern and produce antibodies for it in the future. (To learn more about this process, check out these articles from the CDC and Civil Beat.)

How is this important? Not only is this method safer than previous vaccines (mRNA does not penetrate the nuclear membrane, so it doesn’t modify DNA and is incapable of infection because it is only a tiny piece of the full virus), it is applicable to any virus with a recognizable antigen that can be used to instigate the production of antibodies. It has also been studied for decades before being applied to the SARS-CoV-2 virus, so a large amount of information is available on its potential applications. Influenza, Zika, rabies, and cytomegalovirus mRNA vaccines are already in their early-state clinical trials, and there are studies that suggest mRNA is on its way to curing cancer next.

Generative Pre-Trained Transformer 3 (GPT-3)

GPT-3 is currently the most advanced language model to produce human-like text. While it doesn’t exactly pass the Turing test, it is exemplary at communicating simple facts in a grammatically accurate manner, following directions, and occasionally dabbling in psychoanalysis, resulting in fascinating conversations about topics ranging from veganism to the apocalypse. The model has resurfaced major concerns about its ability to mass-produce fake news articles (such as this one) and generate extremely realistic text.

Its creation represents a major step forward in the area of generative adversarial networks (GANs), which learn to make increasingly more convincing text or images from a discriminator model. The discriminator model simultaneously learns how to distinguish between a GAN’s input versus a real input (in the case of GPT-3, the discriminator distinguishes between the generator’s text and a real text sample scraped from the web). GANs are an advancement towards general AI, or AI that no longer is designed solely to perform a single task, but rather a wide range of tasks that it was not trained to do. GPT-3 is capable of answering questions it was not trained to answer; for example, the model was never trained to do arithmetic, yet it can answer fairly complex mathematical questions. Because of the innate architecture of GANs, GPT-3 is incapable of “memorizing” data, because it never directly interacts with text samples, making its responses truly computer generated. GANs also provide an interesting look into our own methods of learning: similar to a human child, GANs produce data (in humans, we would start speaking or writing), and they are told by a discriminator (in humans, an adult figure like a teacher or parent) what they did wrong, which they can use to learn and modify their behavior.

Virtual Reality (VR) Technology

With the buzz around this concept of the metaverse, it is easy to pass off Meta (previously Facebook) and other tech giants’ attempts at it as marketing schemes. And yet, when trying out VR for the first time, it’s impossible to deny the inevitably powerful (and potentially dangerous) implications the technology has. There are significant ethical questions that need to be raised about the line between the real and the virtual, and, while VR has become wildly successful, it will likely be a long time before it shapes the metaverse. Much like general AI, it is likely that the metaverse will not be realized for a long time, as it will likely take shape as a slow, gradual movement as virtual communication becomes more capable of replicating or improving human interaction.