Exploring the Capabilities of ChatGPT as a use case generation tool (part 1) 

We are excited to introduce our blog series that will explore the potential of ChatGPT to generate use case ideas for artificial intelligence projects across different industries.  

To achieve this, we leverage ChatGPT’s natural language processing capabilities to generate a diverse range of innovative ideas. We won’t modify the content generated by ChatGPT, but we’ll add our own perspectives and comments to enrich the ideas created and give reference to our own experiences.  

Throughout this blog series, we will examine the relevance, creativity, and breadth of ChatGPT-generated use cases. By doing so, we will validate if generative AI could be a brainstorming partner across different topics. We hope it will inspire our readers to explore the potential of AI technologies.  

Generating AI use case ideas for the automotive industry 

In our first attempt, we prompted ChatGPT to generate a blog post about AI use cases in the automotive industry. ChatGPT generated the following response: 

“Artificial intelligence (AI) has become a buzzword across many industries, including the automotive industry. AI is being used in various areas of the automotive sector, from manufacturing to marketing. In this article, we will explore the use cases of AI in automotive with a breakdown of four key technologies: computer vision, text analytics and NLP, machine learning, and recommendation engines. 

Computer Vision 

Computer vision refers to the ability of machines to interpret and understand visual data from the world. It is a crucial technology for autonomous vehicles, as it enables the vehicle to perceive the environment and make decisions accordingly. Here are some examples of how computer vision is being used in the automotive industry: 

1. Lane departure warning: Computer vision technology is used to detect when a vehicle is veering out of its lane and alert the driver to take corrective action. 

2. Pedestrian detection: Computer vision is used to detect pedestrians on the road and warn the driver if there is a risk of collision. 

3. Object recognition: Computer vision can recognise objects on the road, such as traffic lights, road signs, and other vehicles, to help the vehicle make decisions about speed and direction. 

4. Night vision: Computer vision can be used to provide enhanced visibility in low light conditions, making it easier for drivers to see the road ahead. 

5. Gesture recognition: Computer vision can recognise hand gestures made by drivers, allowing them to control the vehicle’s infotainment system without taking their hands off the wheel. 

Text Analytics and NLP 

Text analytics and natural language processing (NLP) refer to the ability of machines to understand and interpret human language. Here are some examples of how text analytics and NLP are being used in the automotive industry: 

1. Sentiment analysis: Text analytics can be used to analyse customer feedback on social media, allowing automakers to understand customer sentiment and make improvements accordingly. 

2. Voice assistants: NLP is used in voice assistants, such as Apple’s Siri and Amazon’s Alexa, which can be integrated into vehicles to provide drivers with information, entertainment, and assistance. 

3. Chatbots: NLP can be used to develop chatbots that can interact with customers, providing them with information about vehicles, services, and pricing. 

4. Natural language search: NLP can be used to enable drivers to search for information using natural language, such as “find the nearest gas station.” 

Machine Learning 

Machine learning refers to the ability of machines to learn from data and improve their performance over time. Here are some examples of how machine learning is being used in the automotive industry: 

1. Predictive maintenance: Machine learning algorithms can be used to predict when a vehicle is likely to require maintenance, reducing downtime and improving efficiency. 

2. Autonomous driving: Machine learning is a crucial technology for autonomous vehicles, enabling them to learn from their environment and make decisions accordingly. 

3. Personalised marketing: Machine learning can be used to analyse customer data, allowing automakers to deliver personalised marketing messages and offers. 

4. Quality control: Machine learning can be used to analyze data from sensors and cameras on the production line, enabling automakers to identify and fix quality issues in real-time. 

Recommendation Engines 

Recommendation engines are AI algorithms that analyze user data to recommend products, services, and content. Here are some examples of how recommendation engines are being used in the automotive industry: 

1. Personalised car configuration: Recommendation engines can be used to analyse customer data and recommend personalized configurations for new cars, such as color, trim, and accessories. 

2. In-car entertainment: Recommendation engines can analyse user data to recommend music, podcasts, and other content that drivers and passengers are likely to enjoy. 

3. Service recommendations: Recommendation engines can analyse vehicle data to recommend maintenance services, such as oil changes, tire rotations, and brake inspections. 

4. Predictive maintenance: Recommendation engines can be used to analyse vehicle data, such as sensor readings and driving patterns, to predict when maintenance is needed. This can help automakers proactively schedule maintenance appointments, reducing downtime and improving efficiency. 

Overall, AI technologies are revolutionising the automotive industry, enabling vehicles to be smarter, safer, and more efficient. From autonomous driving to personalised marketing, AI is transforming the way automakers operate and deliver value to their customers. As the technology continues to evolve, we can expect to see even more innovative use cases in the future.”

Our commentary 

ChatGPT gives 4-5 possible uses of AI across computer vision, NLP and text analytics, machine learning and recommendation engines in the automotive industry. Most of them are quite popular and any experienced AI specialist could generate similar ideas. We could add several other use cases for each of the technologies. 

For computer vision use cases, ChatGPT could also mention the automated inventory management and monitoring systems (they fall under the object recognition use case). Such systems could be leveraged to generate near real-time HD Maps of road assets such as traffic signs, and lights, light and electricity poles and manholes. It could also be used to determine road surface quality, thus allowing for more proactive road network management.  

We’ve implemented such systems for our automotive and geospatial clients. If you want to learn more about how such systems could be built using modern deep-learning-based computer vision models, see our blog post: Object detection in large panorama images. 

Another use case would be a driver monitoring system, that will monitor whether the driver pays attention while driving. We’ve implemented such a solution for one of our automotive clients, and you can learn more about it here: How we built a Driver Status Monitoring system. 

Regarding NLP, it is interesting to see that ChatGPT does not mention any use of LLMs. Perhaps because it was trained on data up to 2021 and is not aware of the recent revolution. We could easily extend the list of possible use cases by mentioning smart assistants (ChatGPT itself included). Those assistants could enhance the user experience by being fine-tuned on car manuals, organisation FAQs, internal documentation and other text sources.  

Asking questions instead of querying keywords would have been a much nicer experience. It also mentions sentiment analysis, which is crucial for better customer relationship management across many (link: Customer Sentiment Analysis: a way to refine your business strategy).  

For the Machine Learning use cases section, the proposed cases are standard across different industries. ChatGPT mentions personalised marketing, which would fit better under the recommendation engines section. We could prompt ChatGPT to come up with more unique use cases. We could, for example, limit it to things related to the interior of cars or to things related to the supply chain. The quality control use case is interesting as we have been involved in projects like that. By leveraging AI technologies to monitor quality, companies can improve their processes and product quality, generate cost savings and improve productivity. 

For the recommendation engine section, we see that the last of the four proposed use cases can be solved with classic ML techniques – the predictive maintenance use case. If you want to learn more about what we can do with the recommendation engines see our offering page: Recommendation engines. 

Out-of-the-box ideas 

We feel like most of the use cases that ChatGPT generated were quite common and experts would generate and describe them quite quickly. To get something more unusual, we can further prompt it to give us more “out-of-the-box” kind of ideas to challenge its creativity.  

ChatGPT response: 

“Sure, here are some out-of-the-box examples of AI technologies in the automotive industry that could have a significant impact: 

Computer vision: 

• Real-time driver monitoring: Computer vision can be used to track the driver’s facial expressions, eye movements, and body posture to detect signs of fatigue, distraction, or impairment. This could alert the driver or autonomous vehicle system to take corrective action, reducing the risk of accidents. 

• Vehicle customisation: Computer vision could be used to personalise the vehicle’s interior based on the driver’s preferences. For example, the system could detect the driver’s height and adjust the seat, mirrors, and steering wheel accordingly. 

Text analytics and NLP: 

• Customer emotion detection: Text analytics and NLP could be used to detect the emotional state of customers interacting with the dealership or service centre. This could allow dealerships to address concerns in real-time, improving customer satisfaction. 

• In-car communication: NLP could be used to enable drivers to communicate with passengers in different languages. The system could translate the driver’s speech and display it on a screen or play it through the car’s speakers. 

Machine learning: 

• Traffic prediction: Machine learning could be used to predict traffic conditions in real-time and suggest alternative routes to the driver. This could reduce congestion and improve travel times. 

• Health monitoring: Machine learning could be used to monitor the health of the vehicle’s components, such as the engine, battery, and brakes, and alert the driver or service centre when maintenance is needed. 

Recommendation engines: 

• Energy optimisation: Recommendation engines could be used to optimise the vehicle’s energy consumption based on the driver’s habits, road conditions, and weather. This could help to reduce emissions and save fuel. 

• Vehicle sharing: Recommendation engines could be used to match car owners with potential renters based on their location, availability, and preferences. This could enable car owners to monetise their idle assets and reduce the number of cars on the road.” 

Now, for Computer Vision and Machine Learning, ChatGPT mentions an example we have already deployed in Spyrosoft – a driver monitoring solution.  

The NLP section now has more unique use cases. NLP could be used here with the assistance of the voice-to-text models such as OpenAI’s Whisper.  

The ML section is interesting too, supply management and defect detection use cases can leverage the ML algorithms and techniques to optimise the efficiency of operations and production costs. If you want to learn more about the first one, see our article “Defect Detection Management – how to do it effectively”.  

The recommendation engine use cases are good. The energy optimisation recommendations are a great example. Such a feature could be implemented by leveraging next-best-action solutions or reinforcement learning techniques.  

Summary  

ChatGPT can generate use case ideas for AI projects across different technologies. We can use it as a brainstorming and problem-solving companion. It provides general information for high-level questions, but we can dig deeper into selected topics to get more details. We can also encourage it to generate “out-of-the-box” ideas if we are unhappy with the current results.  

In the article, we added our own commentary and insights to assess use case relevance, creativity, and breadth. While not all suggestions were a perfect fit, many showcased promising opportunities for further exploration and development.  

This post serves as a testament to the ever-evolving world of AI and its growing influence on various industries and applications. Stay tuned for our upcoming posts, where we dive deeper into this exciting topic. 

If you’re interested in cooperating with us, visit our AI and Machine Learning offering page or contact us directly via the form below.