Human-Centered Design

Blog

ChatGPT and the Future of Artificial Intelligence
Reading time: about 8 min

ChatGPT is one application that leverages OpenAI’s GPT language prediction model. As of this writing, the GPT-4 version recently rolled out, boasting “human-level performance.” But rather than telling you about this manifestation of an artificial intelligence solution based on my own research, let me conduct an interview with ChatGPT, leveraging its earlier GPT-3 engine. But to make this a bit more entertaining, let’s rename this AI entity “Wall-E”, the personified robot from Pixar’s movie of the same name. The following discussion is taken directly from ChatGPT’s responses to my questions: 


Rob Fay: What is a basic overview of OpenAI and its language prediction model? 

Wall-E: OpenAI is an artificial intelligence research laboratory consisting of a team of researchers and engineers dedicated to creating advanced AI systems. It was founded in 2015 by a group of tech luminaries, including Elon Musk, Sam Altman, Greg Brockman, Ilya Sutskever, and others. 

One of OpenAI's most well-known and groundbreaking projects is its language prediction model, which is also known as GPT (Generative Pre-trained Transformer). GPT is a deep learning model that uses a neural network to analyze and understand natural language text. It can generate human-like text by predicting the next word or phrase based on the context of the preceding text. 

GPT is trained on a massive dataset of text from the internet, including books, articles, and websites. The model uses unsupervised learning to learn the patterns and structures of language, allowing it to generate coherent and realistic text without any specific instructions or examples. 

GPT has been used for a variety of applications, including language translation, question-answering systems, chatbots, and more. It has set numerous records in natural language processing, such as achieving state-of-the-art results in language modeling, language generation, and machine reading comprehension. 

Rob Fay: How can artificial intelligence help the Centers for Medicare and Medicaid Services fulfill its mission? 

Wall-E: Artificial intelligence (AI) can help the Centers for Medicare and Medicaid Services (CMS) fulfill its mission in several ways: 

  1. Fraud detection: CMS can use AI to detect fraudulent activities by analyzing large volumes of data to identify unusual patterns and behaviors. This can help prevent improper payments and save money.
  2. Predictive analytics: AI can be used to predict health outcomes and identify patients who are at risk for certain conditions. This can help CMS develop more targeted and effective programs and interventions to improve health outcomes.
  3. Claims processing: AI can help CMS process claims more efficiently by automating certain tasks, such as data entry and verification. This can reduce errors and improve the speed and accuracy of claim processing.
  4. Resource allocation: AI can help CMS optimize resource allocation by analyzing data on patient needs and service utilization. This can help CMS allocate resources more effectively and efficiently, ensuring that patients receive the care they need.
  5. Personalized medicine: AI can help CMS develop personalized treatment plans for patients by analyzing patient data and identifying patterns and trends. This can help improve health outcomes and reduce healthcare costs by tailoring treatment plans to individual patient needs. 

Overall, AI has the potential to help CMS fulfill its mission by improving the efficiency, effectiveness, and quality of healthcare services. 

Rob Fay: How can ChatGPT help the Centers for Medicare and Medicaid Services fulfill its mission? 

Wall-E: As an AI language model, ChatGPT can help the Centers for Medicare and Medicaid Services (CMS) fulfill its mission in several ways: 

  1. Information dissemination: ChatGPT can provide accurate and up-to-date information about CMS programs and services to beneficiaries, providers, and stakeholders. This can help improve access to information and promote transparency.
  2. Education and outreach: ChatGPT can help CMS educate beneficiaries and providers about healthcare topics, such as preventive care, chronic disease management, and healthy behaviors. This can help promote better health outcomes and reduce healthcare costs.
  3. Customer service: ChatGPT can assist beneficiaries and providers with their inquiries and concerns about CMS programs and services. This can help improve customer satisfaction and reduce call center wait times.
  4. Data analysis: ChatGPT can analyze large volumes of data to identify trends and patterns that can inform policy decisions and program development. This can help CMS develop more effective and efficient healthcare programs and services.
  5. Quality improvement: ChatGPT can provide feedback and insights to CMS about the quality of healthcare services and the patient experience. This can help CMS identify areas for improvement and develop strategies to enhance the quality of care. 

Overall, ChatGPT has the potential to assist CMS in achieving its mission by providing information, education, customer service, data analysis, and quality improvement support. 

Rob Fay: What technical positions could be replaced by AI in the future? 

Wall-E: Artificial Intelligence (AI) has the potential to automate many tasks and job functions that are currently performed by humans. Here are a few examples of technical positions that could be replaced by AI in the future: 

  1. Data entry and processing: AI can be used to automate many data-related tasks, including data entry, data processing, and data analysis. This could reduce the need for human workers to perform these tasks.
  2. Quality assurance and testing: AI can be used to automate software testing and quality assurance tasks, which could reduce the need for human testers and improve the speed and accuracy of testing.
  3. Technical support and customer service: AI-powered chatbots and virtual assistants can be used to provide technical support and customer service, which could reduce the need for human support agents.
  4. Cybersecurity: AI can be used to automate cybersecurity tasks, including threat detection and response, which could reduce the need for human cybersecurity professionals.
  5. Manufacturing and assembly line jobs: AI-powered robots and automation systems can be used to perform manufacturing and assembly line tasks, which could reduce the need for human workers. 

It's important to note that while AI has the potential to automate many tasks and job functions, it is unlikely to completely replace human workers in the near future. Instead, it is more likely that AI will be used to augment and enhance human capabilities, rather than replace them. 

Rob Fay: What are the criticisms of ChatGPT about its biases? 

Wall-E: One of the main criticisms of ChatGPT and other language models is that they can exhibit biases based on the data they were trained on. This is because language models learn from large datasets of text, and if these datasets contain biases or stereotypes, the model may learn and replicate those biases in its output. 

Here are some specific criticisms of ChatGPT regarding biases: 

  1. Gender and racial biases: Some studies have found that ChatGPT and other language models can exhibit gender and racial biases in their output. For example, the language model may associate certain professions or activities with specific genders or races, which can perpetuate stereotypes and reinforce existing biases.
  2. Lack of diversity in training data: Another criticism of ChatGPT is that it may not be trained on diverse datasets that represent a wide range of languages, cultures, and perspectives. This can result in biases and inaccuracies in its output when used in contexts that it was not specifically trained for.
  3. Amplifying harmful content: ChatGPT has also been criticized for its ability to generate and amplify harmful content, such as hate speech or misinformation. This can have negative effects on individuals and society as a whole.
  4. Ethics and transparency: Some critics have also raised concerns about the ethical implications of using ChatGPT and other language models, and have called for greater transparency in how these models are developed, tested, and deployed. 

It's important to note that OpenAI, the organization behind ChatGPT, has acknowledged these criticisms and has taken steps to address them, such as releasing more diverse training datasets and working to reduce biases in its models. However, the issue of biases in AI is complex and ongoing, and there is still much work to be done to ensure that language models like ChatGPT are as fair and accurate as possible. 

Rob Fay: Thank you Wall-E, for taking the time (average answer took ~10 seconds) to answer our questions. 

 If you’d like to learn more about OpenAI or experiment for yourself with ChatGPT, please refer to the following: 



A head shot of Rob Fay

ROB FAY
Rob currently leads the CCSQ Human-Centered Design Center of Excellence (HCD CoE). The HCD CoE is an organization that impacts the way the CCSQ delivers policy, products and services to its customers. Through the provision of education, support and resources, he promotes the continued implementation and usage of HCD best practices and seeks to fulfill the charge of OMB Circular A-11 Section 280 (i.e., “Managing Customer Experience and Service Delivery”). For over 20 years Rob has focused on making products and services delightful and easy to use by leading research and design initiatives at government agencies like CMS, NIH, and USPTO and commercial organizations including Blackboard and Allegis Group. Rob holds a Master of Information Management and a Master of Science in Marriage & Family Therapy from the University of Maryland, College Park.


  • No labels