This past decade has seen the birth of many AI technologies invading various sectors. Medicine and healthcare are having their fair share of this wave. Just at the point where we are yet to harness all, there is to AI in healthcare, the world stands in awe of a new AI technology. This new technology known as GPT-3 stands tall with the best language models and is better than prior models. It is also interesting to note that this technological advancement is the very first of its kind. Created by OpenAI “an artificial intelligence research laboratory based in San Francisco, GPT-3 is better at creating superb contents. Although there are several concerns about its use in healthcare, yet, its future in medicine looks promising.
What is GPT-3?
GPT-3 stands for Generative Pre-trained Transformer 3. The language is based on 175 billion parameters. It uses pre-trained data and improved algorithms to generate flawless-top-notch-text and commands. It achieves this great feat by sourcing for a larger percentage of its pre-trained dataset from a filtered version of the dataset called Common Crawl along with text from Wikipedia and its parent company “ OpenAI. For every command given, it comes up with the best possible response.
How GPT-3 Works
GPT-3 works by taking a piece of language as input and producing the most useful language text as output. It does this using a large number of available datasets online. OpenAI has made available a wide array of resources for GPT-3 to understand how languages work. It studies the meaning of words and their various contexts to generate its text. One unique feature that makes this AI technology stand out from the rest is its weighting process. Its ability to hold 175 billion in its memory to process a single query makes it the best artificial neural network ever made.
The GPT-3 Ability
Because GPT-3 can create and translate anything that has a language structure, in the nearest future, it will generate interesting health articles that people will love to read by taking advantage of Google search algorithms and the many data available online. It will also summarize long-form medical articles into understandable ones and even create computer codes that will be beneficial to healthcare. There are already predictions that it could have big implications on the development of future health apps and software. Although before June 2020, very few developers had access to the codes created by GPT-3 through an Application Programming Interface maintained by OpenAI. Ever since it was made public it has generated positive reviews for its content-generating ability.
GPT-3 in Healthcare
GPT-3 can handle administrative tasks like appointment booking successfully. It can help medical professionals pull out vital information from long paper works. For instance, finding a patient’s insurance information before a medical examination. With this, nurses and doctors do not have to go through the arduous task of scanning through large paper files or online documentations to get vital details.
COVID-19 has caused a shift in people’s mental health, creating new mental health concerns. While many people have become more anxious due to the ongoing pandemic, Steven Taylor “ a professor in psychiatry at the University of British Columbia debates that life will not remain the same again for 10 – 15% of people because of its adverse effect on their mental wellbeing. It is estimated that many people will continue to adopt mental health apps to meet their mental health needs.
Hence, one promising area of healthcare where GPT-3 can be successfully applied is the mental health space. GPT-3 will listen to your mental health concerns and come up with temporary actionable tips to make you feel better pending a doctor’s appointment. It can suggest you take a walk, go see a friend, or get involved in a charitable cause. At the moment. Doc.ai “ a digital health company based in Palo Alto, is exploring the power of GPT-3 in mental health.
Future Interventions Amidst Growing Concerns
The innovators of GPT-3 have warned that it presently lacks the medical and scientific wherewithal to make advanced medical documentation, diagnosis, and prescription. This is because there is more to making clinical inferences than taking advantage of a large number of online data. In some scenarios, while it gave accurate answers, at other times, it gave wrong recommendations and if ignored, these contradictions can be fatal.
GTP-3 is at its early stage just as AI and robotics were “ once upon a time, overtime its algorithms will undergo lots of fine-tuning to perfectly fit into healthcare. In the nearest future, it will take on the role of a digital physician assistant accurately diagnosing and following up on patient care. Future physicians may not have to see patients except it is very important and till then, cheers to the latest AI tech in healthcare.
