202309-thumbnail_Cover Story

ChatGPT engineer or engineer ChatGPT?

By Hong Kong Economic Times

If you choose to listen to this article, you are welcome to download the PDF version of the Journal (September 2023 issue) and activate the "Read Out Loud" function in Adobe Reader. For more details, please read the user's note.

 

Many people already know about ‘ChatGPT’, the powerful new generative artificial intelligence (AI) system that debuted in November 2022, also known as a ‘Large Language Model’ (LLM), designed to generate human-like text content by predicting word sequences based on its training data.

 

Self Photos / Files - photo2

The rise of artificial intelligence (AI) and natural language processing (NLP) has led to the development of various chatbot technologies

 

Definition of generative AI and LLM

 

‘Generative AI’ is a broad term that can be used for any AI system whose primary function is to generate content1, be it text, image, audio, video, code, and more. In most cases, generative AI is prompted by a user (referred to as a ‘prompt’ in the AI world), and then the AI attempts to create something that matches the given description.

 

In addition to ChatGPT, there are other examples of generative AI systems including image generators such as ‘Midjourney’ or ‘Stable Diffusion’, code generation tools such as ‘Copilot’, or audio generation tools such as ‘VALL-E’ or ‘resemble.ai’.

 

In the AI world, the term ‘LLM’ refers to systems that work with language. LLM like ChatGPT are currently trained on vast collections of articles, Wikipedia entries, books, internet resources, and other inputs to generate human-like responses to natural language queries.

 

In the months since its debut, ChatGPT has become a global phenomenon, with millions of people all over the world using it. This global influence of ChatGPT is reflected in user data.

 

ChatGPT was the ‘fastest-growing consumer application in history,’ according to research from UBS.2 It reached 100 million users just two months after its launch.3 In comparison, it took Instagram approximately 2.5 years after its global launch to reach 100 million users.4 However, after Meta’s new social media app ‘Threads’ launched on 5 July, it achieved the milestone of 100 million users within five days5. Threads’ rapid adoption, however, is based on Facebook's extensive user base. In contrast, ChatGPT emerged independently and established its user community from scratch. As a result, ChatGPT remains distinctive in terms of global influence.

 

Self Photos / Files - Screen Shot 2023-09-01 at 11.46.37 AM

OpenAI has become synonymous with the generative AI boom

 

Despite reports of its convincing yet sometimes inaccurate results, ChatGPT is considered by experts to be the future of most workplaces.

 

For instance, a recent report by the Organisation for Economic Co-operation and Development (OECD)6, a 38-member bloc mainly comprising wealthy nations, revealed that over a quarter of jobs in the OECD rely on skills that could be easily automated in the upcoming artificial intelligence revolution. According to the OECD, workers are concerned about potential loss of their jobs to AI. The survey was conducted before the rapid rise of ChatGPT.

 

According to Reuters7, ChatGPT recently passed law exams in four courses at the University of Minnesota, which is ranked as the 21st best law school in the country by U.S. News & World Report. A group of four law professors at the university utilised the popular AI chatbot to generate answers for exams in four courses during the last semester. Then they blindly graded these answers alongside actual students’ tests. In terms of results, ChatGPT achieved an average performance of C+, which was lower than the humans’ B+ average. Nonetheless, if this grade was applied across the entire curriculum, the chatbot would still earn a law degree.

 

In the medical field, ChatGPT has answered US medical licensing exam questions accurately over 90% of the time8. It has even outperformed the examination skills of some specialised licensed doctors. The researchers in charge of the experiment mentioned that ChatGPT could translate discharge information for a Portuguese-speaking patient and simplify technical jargon into language that a 6th grader could easily understand.

 

As ChatGPT becomes increasingly popular and continues to achieve remarkable results in professional exams, there is a growing anticipation that its influence will extend to virtually every kind of workspace. Some people are beginning to wonder: Does ChatGPT have the potential to make a significant impact in the field of engineering? To be more precise, could it eventually replace specific job roles in the industry?

 

ChatGPT as software engineer?

 

It is not surprising that programming is a professional domain where AI performs exceptionally well. According to CNBC9, ChatGPT has successfully passed Google's coding interview for level 3 software engineers. Google conducted this experiment as part of their recent assessment of multiple AI chatbots that they are considering adding to their site. In fact, level 3 is considered an entry-level position on the engineering team at Google. The average total compensation for this job is approximately US$183,000 (or HK$1,400,000) per year.

 

Self Photos / Files - Screen Shot 2023-09-01 at 11.53.06 AM

Programming is a professional field where AI performs exceptionally well

 

However, generative AI tools like ChatGPT are not meant to replicate the experiences, creativity, critical thinking, and problem-solving abilities of a real-life software engineer. Rather, ChatGPT is intended to assist with certain tasks, and it requires human supervision and direction to operate effectively.

 

AI turns everyone into a designer

 

‘Text to Image’ design is another remarkable achievement of AI. The images generated by AI have reached a level where distinguishing between real and fake becomes challenging.

 

AI-assisted design tools can certainly help democratise the design process, making it more accessible to a wider range of people and potentially leading to new and innovative approaches to design. The concept behind this statement is that AI-powered design tools can automate numerous processes integral to design creation, including tasks such as selecting colour schemes, choosing fonts, and optimising layouts. This would make it easier for individuals without a design background to create professional-looking designs, without needing the same level of expertise as a traditional designer.

 

However, while AI can certainly aid in design, it is unlikely that it will completely replace human designers. Many other aspects of design require creativity, intuition, and a profound grasp of user experience - qualities that remain challenging for AI to mimic.

 

Self Photos / Files - Screen Shot 2023-09-01 at 11.55.26 AM

AI image generator tools will get you the best AI-generated images using text-to-image AI

 

For engineers whose work involves both programming and design, the impact of AI is significant, encompassing both positive and negative effects. An engineer’s tasks often require a combination of programming and design skills, which can be substantially enhanced by the influence of AI. In fact, in many engineering fields, the use of AI has become increasingly necessary, equipping engineers with powerful tools to streamline their work and improve the efficiency of their designs. With AI, engineers can perform complex simulations and calculations in a matter of seconds. Additionally, AI can help engineers to identify potential issues and offer valuable insights, resulting in more effective and innovative solutions.

 

It is clear that the impact of AI on engineering cannot be overstated, and we expect to see more exciting and groundbreaking developments in the future. It is essential for engineers to possess a comprehensive understanding of the processes that AI has the potential to replace. However, it is crucial to note that the term ‘replace’ in this context does not necessarily imply the complete automation of human tasks by AI. Instead, AI can be harnessed to make work easier and enhance efficiency.

 

For instance, ChatGPT has the potential to become a valuable tool by streamlining communication and problem-solving between engineers and other stakeholders. It can automate routine tasks and offer support in non-technical areas, allowing engineers to focus on more challenging and complex tasks.

 

The development of AI has opened up new possibilities for engineers to streamline operations, enhance productivity, and achieve greater accuracy. Therefore, engineers should not only be aware of the of AI’s potential to replace certain processes, but explore ways to integrate AI into existing systems to optimise performance and create new opportunities.

 

Engineer x ChatGPT

 

Engineering is a complex field that involves a wide range of disciplines, including mechanical, electrical and civil engineering. In fact, the adoption of AI is not a new phenomenon in the building Electrical and Mechanical (E&M) industry. The HKSAR Government, through its Electrical and Mechanical Services Department (EMSD), has identified Semantic AI as an important component in developing knowledge-based systems, including the traditional building E&M industry that can increase efficiency and promote sustainable solutions in smart buildings.10

 

EMSD has developed an award-winning Semantic AI model that enhances the engineering facility in buildings, ultimately conserving energy. With a portfolio of more than 8,000 facilities, EMSD established its first Regional Digital Control Centre (RDCC) as a centralised data processing and modelling hub. The centre utilises an open data platform and AI solution, collecting up to 600,000 data points from a single building every day. The system is supported by an analytic engine with AI and machine learning capabilities, which is used to effectively manage and improve the energy performance of its building portfolio.

 

In addition to the government's efforts, tech startups also play a significant role in addressing long-standing industry-specific challenges such as an ageing workforce, safety incidents, and declining productivity. The winning team of the HKUST-Sino One Million Dollar Entrepreneurship Competition founded AutoSafe, a start-up that leverages AI technology to modernise and automate worksite safety monitoring while improving efficiency.11

 

AutoSafe's AI-driven construction site monitoring system uses cameras and software to prevent accidents. The multi-camera tracking system collects walking trajectories efficiently from various locations on a construction site and alerts workers in close proximity to heavy equipment. This proprietary system has attracted the interest of potential users including the Drainage Services Department, a major construction contractor, and a large real estate developer in the Mainland.

 

The engineering profession is familiar with repetitive tasks. Beyond the current application scenarios, ChatGPT has the potential to automate certain aspects of basic and standard tasks.

 

It can be a valuable tool for engineers as it can assist them in carrying out routine tasks more efficiently and accurately, allowing them to focus on more complex and creative aspects of their work.

 

Furthermore, ChatGPT aids engineers in computations, generating engineering-based text, supporting design modelling in the early stages, producing reports, translating, and offering answers to questions. Furthermore, it possesses the capability to learn through interactions with engineers, gradually becoming more valuable over time. Therefore, this technology is gaining favour within the engineering field and across its various subcategories.

 

Mechanical engineering

 

Mechanical engineers design power-producing machines such as electric generators, internal combustion engines, and steam and gas turbines. These engineers can use ChatGPT to help generate 3D models to simulate physical systems, produce technical reports and documents, perform failure analysis, participate in product development, conduct risk management, expedite design processes, ensure designs are safe and reliable, and free up mechanical engineers for higher-value tasks.

 

Electrical engineering

 

Electrical engineers design, develop, and test electrical equipment such as electric motors, radar and navigation systems, communication systems, and power generation equipment. Like other engineering fields, electrical engineers can utilise ChatGPT to answer technical questions and explain complex concepts and theories in simple terms. Additionally, the chatbot can provide suggestions and advice on electrical engineering solutions, share research and industry insights, and discuss emerging technologies.

 

Civil engineering

 

Civil engineers are responsible for brainstorming, designing, building, and maintaining infrastructure projects in both the public and private sectors. These projects include roads, buildings, airports, bridges, tunnels, dams, and systems for water supply and sewage treatment.

 

In the construction industry, civil engineers are already using ChatGPT for various tasks. The chatbot can answer civil and construction questions, offer technical support for projects and best practices, provide education and training, and streamline communications across projects.

 

Furthermore, ChatGPT can provide support in project designs and structural analysis. Engineers first input building specifications and materials, then ChatGPT provides suggestions on or insights into potential vulnerabilities. The chatbot can also aid in creating Building Information Modeling (BIM) information, which is utilised in construction projects to manage and share project data. With ChatGPT and other generative AI tools, users can generate images and videos using text input parameters. This makes it possible to create informative content, such as videos on building information modeling (BIM).12

 

By incorporating AI-based design exploration, engineers can improve their designs using data from simulations, models, and past projects. They can also make precise changes across all design areas.

 

A crucial element is mastering the skill of posing impactful questions. This requires comprehensive understanding of the situations, breaking complex problems into smaller parts, and fine-tuning inquiries. In other words, the ‘why’ questions carry greater significance than the ‘how’ questions. ChatGPT serves as a great resource for teaching, learning and data compilation. However, when seeking straightforward ‘why’ answers, the responses might lack a robust theoretical explanation supported by evidence. Engineers should further explore and ask more specific questions for obtaining comprehensive insights.

 

These are just a handful of examples of ChatGPT’s utilisation in the engineering field, marking only the beginning. However, advancements in AI have also led to numerous skeptical viewpoints.

 

Examining the flip side, there exist potential risks and challenges linked to the utilisation of ChatGPT in engineering.

 

For example, ChatGPT may not fully understand the nuances and complexities of engineering problems, leading to suboptimal solutions or errors. Additionally, ChatGPT may reinforce existing biases or assumptions in engineering processes and practices, rather than questioning them.

 

To effectively address these potential risks and challenges, it is crucial to approach the application of ChatGPT in engineering with a critical mindset. This means assessing the possible benefits and drawbacks of using ChatGPT in different contexts and ensuring that its applications align with ethical and human-centred principles.

 

Self Photos / Files - photo3

Civil engineers can use AI-assisted BIM tools to make it easier to create and design more accurate 3D models before starting construction

 

Local universities shared their perspectives on AI

 

As the training ground for engineering professionals, colleges may be where changes in collaboration between humans and artificial intelligence begin. When ChatGPT first became a popular topic, there was a lot of discussion in Hong Kong academic circles about its use. Initially, local universities seemed to lean towards prohibiting students from using generative AI tools.

 

However, the situation began to change after The Hong Kong University of Science and Technology (HKUST) granted ‘limited approval’ in early March.13 In a letter to staff, HKUST stated that the decision to allow or disallow the use of AI tools in classes would be left to individual teaching staff. Moreover, starting from 1 June 2023, the university has even developed its own version of ChatGPT for all university members to utilise.14

 

The Education University of Hong Kong became the second university in the city to adopt AI technology. Students are permitted to use AI-enabled tools in their assignments as long as they disclose such usage and ‘articulate their reflections’ on the process.15

 

In August, The University of Hong Kong (HKU) lifted its ban on students using ChatGPT and announced that it would provide a wide range of generative AI tools to both students and teaching staff in the upcoming academic year.16 Unlike other ChatGPT users in the city, HKU students will be able to use the AI application every month without the need for a VPN. Each student will be given a quota of up to 20 prompts every month, along with access to the image generator Dall-E.

 

Conclusion

 

Generative AI is still in its early stages, and its influence is likely to grow as technology continues to advance and improve.

 

For engineers, Generative AI tools such as ChatGPT will be like a calculator to accountants, becoming a basic and necessary tool.

 

However, some people hold a pessimistic view on AI, while others find it challenging to adapt to new technology and are unwilling to start using ChatGPT. In fact, transforming AI into a collaborative work ally goes beyond mere usage; it requires a thorough understanding in order to enable effective control.

 

The sooner this technology is adopted, the sooner engineers can reap its benefits. Therefore, it is recommended that anyone in the engineering field who reads this article starts utilising AI applications and contemplates how to leverage AI to enhance their competitive edge.

 

References

 

  1. Toner H (2023). What Are Generative AI, Large Language Models, and Foundation Models? CSET. [online]. Available at: https://cset.georgetown.edu/article/what-are-generative-ai-large-language-models-and-foundation-models/ [Accessed 24 August 2023]
  2. Hu C (2023). ChatGPT sets record for fastest-growing user base - analyst note. Reuters. [online]. Available at: https://www.reuters.com/technology/chatgpt-sets-record-fastest-growing-user-base-analyst-note-2023-02-01/ [Accessed 24 August 2023]
  3. Milmo D (2023). ChatGPT reaches 100 million users two months after launch. the Guardian [online]. Available at: https://www.theguardian.com/technology/2023/feb/02/chatgpt-100-million-users-open-ai-fastest-growing-app [Accessed 24 August 2023]
  4. World of Engineering (2023) [Twitter]. Available at: https://twitter.com/engineers_feed/status/1621226994670125057 [Accessed 24 August 2023]
  5. Paul K (2023). Meta’s Twitter rival Threads surges to 100 million users faster than ChatGPT. Reuters [online]. Available at: https://www.reuters.com/technology/metas-twitter-rival-threads-hits-100-mln-users-record-five-days-2023-07-10/ [Accessed 24 August 2023]
  6. Reuters (2023). AI puts 27 per cent of jobs at risk of being automated, OECD says, amid rising concerns from workers. South China Morning Post [online]. Available at: https://www.scmp.com/tech/tech-trends/article/3227383/ai-puts-27-cent-jobs-risk-being-automated-oecd-says-amid-rising-concerns-workers [Accessed 24 August 2023]
  7. Sloan K (2023). ChatGPT passes law school exams despite 'mediocre' performance. Reuters [online]. Available at: https://www.reuters.com/legal/transactional/chatgpt-passes-law-school-exams-despite-mediocre-performance-2023-01-25/ [Accessed 24 August 2023]
  8. Brueck H (2023). The newest version of ChatGPT passed the US medical licensing exam with flying colors — and diagnosed a 1 in 100,000 condition in seconds. Insider [online]. Available at: https://www.insider.com/chatgpt-passes-medical-exam-diagnoses-rare-condition-2023-4 [Accessed 24 August 2023]
  9. Dreibelbis E (2023). ChatGPT Passes Google Coding Interview for Level 3 Engineer With $183K Salary. PCMag [online]. Available at: https://www.pcmag.com/news/chatgpt-passes-google-coding-interview-for-level-3-engineer-with-183k-salary [Accessed 24 August 2023]
  10. The Electrical and Mechanical Services Department (2022). A new generation of Artificial Intelligence: Semantic AI on traditional E&M industry. Hong Kong Engineer [online]. Available at: http://www.hkengineer.org.hk/issue/vol50-oct2022/feature_story/?id=17083 [Accessed 24 August 2023]
  11. The Hong Kong University of Science and Technology (2022). AI on the Watch for Worksite Safety. [online]. Available at: https://hkust.edu.hk/news/entrepreneurship/ai-watch-worksite-safety [Accessed 24 August 2023]
  12. Osadchyi O (2023). How ChatGPT Can Transform the AEC Industry. Linkedin [online]. Available at: https://www.linkedin.com/pulse/how-chatgpt-can-transform-aec-industry-oleg-osadchyi/ [Accessed 24 August 2023]
  13. Cruise S (2023). Science and Technology allows students to use ChatGPT on a limited basis. Techgoing [online]. Available at: https://www.techgoing.com/science-and-technology-allows-students-to-use-chatgpt-on-a-limited-basis/ [Accessed 24 August 2023]
  14. Muthann A (2023). HKUST Creates Its Own Version Of ChatGPT For Students, Faculty & Staff. The HK Hub [online]. Available at: https://thehkhub.com/hkust-creates-its-own-version-of-chatgpt-for-students-faculty-staff/ [Accessed 24 August 2023]
  15. Mok L (2023). Hong Kong Education University approves use of ChatGPT in coursework despite bans by two other schools. Hong Kong Free Press [online]. Available at: https://hongkongfp.com/2023/03/24/hong-kong-education-university-approves-use-of-chatgpt-in-coursework-despite-bans-by-two-other-schools/ [Accessed 24 August 2023]
  16. The University of Hong Kong (2023). HKU introduces new policy to fully integrate GenAI in Teaching and Learning. [online]. Available at: https://www.hku.hk/press/press-releases/detail/26434.html [Accessed 24 August 2023]

 

Explore Hong Kong Engineer