ChatGPT: A New Revolution In AI

Greetings from the Dallas AI team in Texas! Dallas AI is starting a periodic newsletter that aims to present opinions on the latest happenings in the analytics space, highlights from our past events as well as share information on upcoming events.

Disclaimer: The newsletters contain thoughts and reactions on technology topics but do not represent the official viewpoints of Dallas AI.

In this February 2023 Newsletter, we present:

• ChatGPT: A New Revolution in AI
• Highlights and snippets from the meetup on "Practical Applications of AI at Scale: Law, Sports & HealthCare.", held in January 2023?
• Information and links to upcoming events in March 2023

ChatGPT: A New Revolution in AI

Ever since OpenAI introduced the world to ChatGPT in November 2022, it has grabbed the world's imagination as well as garnered a significant share of curiosity as well as apprehension. Like so many others technology enthusiasts as well as non-tech folks, we have also invested some time in exploring what it has to offer and our observations have been a mixed bag. On the plus side, it showcased a reasonable level of diligence in writing articles or essays (or even books), answering questions from multiple perspectives in addition analyzing data that is provided to it. On the contrary, tracing the source of the information that ChatGPT provides is a taxing process.

Some of the many ways in which we see ChatGPT providing value:

• For the Society: First and foremost, ChatGPT has opened up a new dimension in how day-to-day business is conducted in the world out there today. It has showcased how communication as well as business transactions can be improved upon. Babar Bhatti highlights how chatGPT can potentially shape our personal as well as the business world in his article guide to ChatGPT for business leaders and executives.
• For us as Individuals: ChatGPT has demonstrated that it can significantly aid in personal productivity by providing reminders, help in maintaining schedules along with offering simple and quick assistance to mundane tasks. Apart from these, ChatGPT has shown the possibility of a logical conversational AI where it can help individuals struggling concentration issues or other anxiety related problems to find simple initial solutions. Of course the ideal solution would be to have a doctor working in conjunction with such a chatbot to help a greater number of individuals, and ChatGPT can be a great starting point. Standard questions that are posed to a patient can be conducted by a system incorporating ChatGPT and hence it can create an initial diagnosis. This creates a win-win situation for the everyone.
• For Developers and the Tech Community: ChatGPT has delineated how coders and developers can reduce the repetitiveness of reusing similar snippets of code and can instead focus their efforts in improving the clarity, conciseness and overall quality of their code. Joseph Ferror has mirrored this sentiment in his article on how ChatGPT could enhance the code.

Even in the financial world, ChatGPT can help in many ways including generating system diagrams.

On the bright side, it does seem ChatGPT can be used in contexts where applications of natural language processing and conversational AI can be instrumental. Its value lies in its versatility!

Although, we just saw a lot of beneficial use cases, there are certain limitations:

While ChatGPT could potentially help or evolve the ways in which people work, there are some obvious pitfalls, which show how we have a long way to go before fully embracing ChatGPT.?

• First, is the method in which It generates answers, doing so by relying on a probability distribution of multiple sources. This effectively means that it is virtually impossible to pinpoint a source as well as truly verify an answer.?
• Second, is the persistent issue of bias in its answers, incorrect representation of facts in the datasets on which the underlying models of ChatGPT are trained on, and such training datasets getting constantly outdated. All these factors bring about a topic that is becoming relevant more than ever, which is the topic of responsible AI.

A lot has been said about responsible AI. Here are a few highlights:

• The level of maturity in AI applications that companies implement in their businesses should be measured more stringently and heavily invested upon.?
• The biggest contributor to responsible AI is transparency. Trust and explainability of AI across processes and functions should be a widespread practice.
• BCG’s article serves as an eye opener on how avoiding risks isn’t a good enough practice anymore and how leaders must invest more in avoiding undesired impacts.
• Accenture’s global research on AI and ethics explains how responsible AI should be adopted and & its large array of benefits.

Hence, it goes without saying that to use ChatGPT responsibly and ethically, a constant monitoring and governance is needed. Proper training and supervision, as well as attention to issues such as bias and security, will ensure that ChatGPT is used in ways that benefits society as a whole.

A Concerning Question: Will ChatGPT change writing forever?

Google programmer Kenneth Goodman tried ChatGPT on a number of exams. It scored 70% on the United States Medical Licensing Examination, 70% on a bar exam for lawyers, nine out of 15 correct on another legal test, the Multistate Professional Responsibility Examination, 78% on New York state's high school chemistry exams multiple choice section, and ranked in the 40th percentile on the Law School Admission Test.?

In the last year or so there have been a number of papers showing that the writings by large language models and chatbots can no longer be differentiated from their human counterparts!

As everything is getting AI-generated, we deserve to know if the content is written by a bot, machine, or human, to maintain authenticity. Edward Tian created GPTZero and its purpose is to tell you if the content is written by a machine like ChatGPT.? Read about Tian’s motivation with GPTZero.??

Insights From Our January 2023 Event

Dallas AI aims on bringing AI meetups every month. The first meetup of 2023 was hosted at Spacee, Inc. featuring Rameez MeeraSahib , CTO of Micro Focus. He discussed the practical side of AI at scale in domains such as law enforcement, healthcare & sports. Here are some thoughts that Rameez shared with us:

1. How do you see AI impacting Law Enforcement, Sports & Healthcare Domains?

AI has already made a significant impact on Law enforcement, Sports, and Healthcare domains, and its impact is expected to grow in the future exponentially.

In Law enforcement, AI is being used to help identify potential suspects, analyze evidence, monitor public spaces and traffic patterns, etc. For example, AI-powered facial recognition technology can help police identify suspects in real-time, and also efficiently manage case investigations where more and more data is in video, audio, and text formats. Instead of spending hours and hours watching a video or audio evidence, AI could do the heavy lifting and find the snippet of the video for final review and use it to accelerate the investigation.

In Sports, AI can help sports teams to make more informed decisions by providing them with real-time data and analysis. AI-powered sensors can track an athlete's performance and provide insights into areas that need improvement, and AI-powered video analysis can help teams identify patterns and improve their strategies. AI is also being used to develop a real-time game strategy. For example, Formula E racing teams are using AI to monitor real-time audio communication between the different teams and identify patterns and anomalies to decide on the next lap strategy.

In Healthcare, AI is being used to improve the accuracy of medical diagnosis and treatment. For example, AI-powered medical coding software can quickly auto-code diagnosis and reduce coding errors.

2. What are your thoughts on the Business impact of AI?

The impact of AI on businesses is significant and far-reaching, and it is changing the way businesses operate in many different ways. One of the main ways that AI is impacting businesses is by improving efficiency and productivity. AI-powered automation and intelligent systems are streamlining processes, reducing the need for human intervention, and enabling businesses to make better use of their resources. This can result in cost savings, increased output, and improved customer experiences. For example, in the healthcare example above, coding errors cost billions of dollars to patients. A single percentage improvement in coding accuracy can save billions of dollars in healthcare spending. By enabling greater efficiency, insights, and innovation, AI is helping businesses to stay competitive and thrive in an increasingly complex and dynamic marketplace.

3. How can ChatGPT be used in Law Enforcement, Sports, and Healthcare?

ChatGPT is an AI application in the early stages of testing and market validation. The success of ChatGPT is an excellent example of how AI can increase productivity in various fields, especially when it comes to research, marketing, business communications, etc. We will see an explosion of such applications of AI in the coming days which will disrupt many industries. In Law Enforcement and Healthcare, the immediate application of regenerative AI technology will be customer service. It can automate the first level of customer support and service using AI instead of humans and reallocate resources to more complex tasks. In sports, marketing departments can use it to accelerate content production to many folds and thereby boosting fan engagement and revenue. ChatGPT is just a start and we are living in a fascinating time where AI will change how we do business and reallocate human resources to innovate and produce better products and services.

Upcoming Events:

We have some great events coming:

1. Artificial General Intelligence (AGI): Are We There Yet?

Achieving Artificial General Intelligence (AGI), i.e., developing intelligent systems whose performance is at par with humans, is the holy grail of AI. Interested in Artificial General Intelligence for real-world scenarios? Dallas AI is hosting Gopal Gupta , Professor of Computer Science at the University of Texas at Dallas and co-director of the Center for Applied AI and Machine Learning, to discuss Large Language Model capabilities and limitations. Click for more details.

2. Machine Teaching: A Systems View of AI Design

Interested in learning about the system-level approaches for Machine Learning Models? Dallas AI is hosting Kence Anderson , AI Researcher, to discuss the implications of humans teaching machines. Join us to learn, share your views, and network with fellow AI practitioners. Click for more details.

3. Generative AI Workshop:

Location: Richardson IQHQ

Dallas AI is offering a one-day workshop on state-of-the-art #Generative #AI. The workshop provides in-depth knowledge and business applications of language models such as ChatGPT and Image generation models such as Dall-E-2 and Stable Diffusion. The workshop includes an introduction, demos, and hands-on projects. Click for more details.

Do you have questions or suggestions for us? Or want to know more about a specific event? Comment below or reach out to us at [email protected].

We highly appreciate suggestions/concerns from the our subscribers, our peers as well as the greater community as a whole.

- The Dallas AI Team