January Observations

From speech-to-text conversion to handwritten text recognition, technology has clearly deciphered most of the modes of human communication. Or has it? After watching University of Technology Sydney Professor Chin-Teng Lin’s TED Talk “The Mind-Reading Potential of AI,” I realized that researchers are currently exploring a whole new means of communication through AI: turning our thoughts and inner speech into text on a screen without the use of a brain implant. 

So, how exactly is this being done? The technology first needs to pick up the brain’s signals through sensors in a cap. Then, two key aspects come into play to help interpret this data. One major component is deep learning. When someone says a specific word, their brain sends out certain signals. After being trained on various examples and large amounts of data, a deep learning model is able to detect patterns in how words and their corresponding signals are related. This means that the model can eventually predict the word a human is thinking about solely based on the brain signals it receives. The other major component is the use of a large language model (LLM). LLMs like ChatGPT are able to differentiate between grammatical and ungrammatical sentences, as well as logical and illogical sentences. Therefore, when a large language model is given the word predictions from the deep learning model, it can account for potential errors in the brain signal decoding process as it formulates the sentences that will appear on the screen. For example, if a predicted word seems entirely out of place, there was likely a mistake earlier in the process, and the LLM could replace the word with a more likely one instead. 

According to Professor Chin-Teng Lin, the technology currently has an accuracy rate of around 50%, and more work is needed to perfect it. However, these advancements are already significant, and an increased accuracy rate in the near future appears probable. It is therefore important to consider the possible implications of this technology. 

The most apparent positive effect is offering a new means of communication for people who are speech impaired. We are all aware of the importance of languages and their ability to build bridges between people and cultures, but some individuals cannot access these benefits as smoothly. Writing thoughts down can be inefficient and time-consuming, so using thought alone would revolutionize communication for people who are unable to rely on verbal communication. 

Unfortunately, every piece of technology also has negative consequences; in this case, the loss of privacy is a major one. When this technology is released to the public, how can we prevent large corporations like social media companies from using it for their own gain? As they start to listen to our private thoughts, they will be aware of our interests and use this to tailor content even more than they already do today. This will not only lead to more echo chambers, but also push consumers to spend money in ways that may not be sustainable in the long term. 

Overall, this technological advance is fascinating. After its harmful impacts are minimized, it can be truly beneficial to numerous individuals.

Credit: James Woodford (NewScientist)

Works Cited

James Woodford. “Mind-Reading AI Can Translate Brainwaves into Written Text.” New Scientist, 12 Dec. 2023, www.newscientist.com/article/2408019-mind-reading-ai-can-translate-brainwaves-into-written-text/. Accessed 30 Jan. 2025.

Lin, Chin-Teng. “The Mind-Reading Potential of AI.” TED Talks, Oct. 2024, www.ted.com/talks/chin_teng_lin_the_mind_reading_potential_of_ai. Accessed 30 Jan. 2025.

Travers, Mark. “A Psychologist Explains What “Mind-Reading” AI Can Do to Us.” Forbes, 22 Nov. 2023, www.forbes.com/sites/traversmark/2023/11/22/a-psychologist-explains-what-mind-reading-ai-can-do-to-us/. Accessed 30 Jan. 2025.