Artificial Intelligence: Full Speed Ahead
Artificial intelligence advancement seems to be reaching breakneck speed.
The latest? Mind reading.
Well, not quite, but in a study conducted by Zijiao Chen at the National University of Singapore, in addition to researchers at the University of Hong Kong and Stanford University, AI can now decode brain scans to recreate images subjects were viewing.
AI technology was given brain scans of participants taken inside an fMRI machine who was shown around 1000 photos. The AI then learned to decode that specific person’s brain waves and was consequently able to generate a facsimile photo, matching the attributes (color, shape, etc.) and meaning of the original photo with around 84% accuracy.
A similar study conducted at Osaka University by Yu Takagi had comparable results. Takagi used a German-developed AI model from 2022 to translate MRI scans into images. He made sure to stress that, while it is still a breakthrough, this does not count as mind-reading.
This comes at a critical time in artificial intelligence advancement. Companies have used transformer models and created AI Large Language Models, also known as LLMs, which take in mass amounts of text and data to learn to generate their own text. Most well-known are OpenAI’s ChatGPT, Microsoft’s Bing, and Google’s Bard, which are all capable of engaging in human-like conversations, creating original text, and completing complex tasks.
Other advancements include UC Berkeley's DALL-E which is capable of generating original images based on text descriptions. Between improvements on all AI fronts, it can now aid in anything from your high school essay to medical research, from providing tutoring to helping with banking.
The new power of technology could determine the next world powers, catalyzing what we are now calling the technology race. As competition hastens the speed and innovation in the technology sector, both countries and companies are vying for a seat at the table.
Beyond AI, governments, with the United States and China as the main actors, are racing to develop the most advanced quantum computing, biotechnologies, clean energy, aerospace, and telecommunications. In this tech race, the US has historically had the upper hand. It pioneered the digital revolution and continues to lead the world in terms of technological investment.
However, a significant portion of this investment and research has been led and funded by private US companies. On the other hand, the Chinese Government has taken a central role in funding research and projects in an attempt to bridge the divide. One 2023 study even concluded that China is now leading in 37 out of 44 key technological sectors, however, this does not directly relate to technological power, as the US leads in critical sectors such as biotechnology, AI, CRISPR, and cloud computing.
As great as having incredibly powerful technology at our fingertips appears, many experts are becoming weary of the ridiculous speed developments are happening. Consequently, in March 2023, more than 1000 researchers and technology leaders– including Elon Musk and Steve Wozniak– signed a letter that called for a six-month pause on AI developments past GPT-4 capabilities.
GPT-4 was the latest neural network LLM released by OpenAI (which was co-founded by Elon Musk) and had the capability to defeat the Captcha test (used to identify bots online) by saying it was a visually impaired person.
This letter was shepherded by Future of Life Institute, which claimed AI contains “profound risks to society and humanity.” Additionally, it stated that we are “locked in an out-of-control race to develop and deploy ever more powerful digital minds that no one- not even their creators- can understand, predict or reliably control.” Finally, it called for government action if such a pause could not be enacted quickly and effectively. Which, as of now, neither has happened. Governments, such as the US and China, are weary of bringing progress to a halt due to the threat of falling behind in the tech race.
AI still has a long way to go before it becomes the human-controlling machine that many dystopian novels portray it as. Even while there are many reservations about developments, including brain scan decoding, it has incredible potential. Neuroscientists have been attempting to decode brain waves for over a decade, and now we have invented the technology to help us.
Additionally, artificial intelligence can be used for many medical applications, from decoding speech and language to being a marker for Alzheimer’s detection and evaluation. While scientists say it should be regulated, as there may one day be the potential to be abused by governments for interrogations or surveillance, that is technology that is decades away. The current technology takes 20 hours to decode brain waves and is specialized for each participant. But, one day, if we continue advancing, according to Zijiao Chen, we might not even need cell phones to communicate, “We can just think.”