NutriNet – a personal assistant for your grocery shopping

18

October

2024

No ratings yet.

Have you ever taken hours browsing around supermarkets searching for the most nutritious food options? Did it take you too much time figuring out which recipes to cook best, with the groceries you bought? Struggles when dealing with food are numerous, starting from choosing products with appropriate nutrients simply not knowing enough recipes.

Therefore, we introduce NutriNet – a mobile application that is going to make shopping and meal planning easier!

• Grocery shopping & meal planning simplification:

NutriNet simplifies grocery shopping and meal planning, by analyzing which products and recipes fit the users’ desired grocery item wishes, nutrition values and store preferences best. NutriNet aims to address the challenges being implicit to personalized nutrition and helps consumers make healthier food choices by simplifying the grocery shopping and meal planning process. This shall be done by eliminating the need to perfectly understand all nutrition values or to search for numerous recipes. NutriNet completes all these tasks for you in real-time and provides clear and accessible recommendations for you.

• Real-time personal assitant: NutriNet acts as a multifunctional application providing value to consumers by solving various food related problems in real-time. Appearing as a chatbot, it is aimed at taking general grocery shopping lists or meal wishes as input query, combined with preferences for food characteristics (e.g., nutrients, allergies) and grocery stores. It then provides brand specific and personalized grocery shopping lists, as well as meal recommendations, if so desired. Moreover, grocery items can also be added into the initial grocery shopping list query, by scanning them with the integrated AR tool. The product will then be detected visually and thus will be integrated into the shopping list input query.

• Long-term customer engagement: NutriNet distinguishes from competition by providing personalized and customized advice. This is possible, as NutriNet consists of a database, having incorporated stock and product information of the major supermarkets in the Netherlands. In contrast, classic applications, which try to meet similar needs (e.g., meal recommendation) rather focus on counting nutrients for the purpose of short-term weight loss, instead of personalizing grocery shopping lists and meal recommendation to enable a healthier lifestyle for users. Those applications are usable on short term but are proven to have low adherence over the time (Chen et al., 2015).

• Personalized recommendations: NutriNet leverages generative AI to offer accurately personalized recommendations. Users can simply enter their preferences, while prompting a grocery list or a meal, such as gluten-free or high in protein, and the generative AI powered application will provide accurately personalized results.
However, personalization and raising awareness for healthy foods are not the only purposes of NutriNet. It also addresses sustainability issues that supermarkets are facing. By gathering consumer purchase and search data in the application, consulting services can be offered to supermarkets, enabling them to plan ordering and stockholding processes more efficient. Hence, supermarkets should be able to reduce food waste due to overstocking on long-term.

NutriNet offers a 2-sided network platform, yielding important value to both, consumers as well as supermarkets.

Contributors

574051 – Duong Dao
728070 – David Wurzer
738898 – David Do
562387 – Roxi Ni

References

Chen, J., Berkman, W., Bardouh, M., Ng, C. Y. K., & Allman-Farinelli, M. (2019). The use of a food logging app in the naturalistic setting fails to provide accurate measurements of nutrients and poses usability challenges. Nutrition, 57, 208-216.

Please rate this

Unlocking Culinary Potential: How tasteTAILOR elevates your Cooking Experience with Generative AI

18

October

2024

No ratings yet.

In today’s fast-paced world, health-conscious individuals and families often struggle to find the time and inspiration to cook nutritious meals. Therefore, we are presenting tasteTAILOR: a revolutionary digital culinary assistant that harnesses the power of Generative AI (GenAI) to streamline your cooking experience. It is designed to create recipes tailored to users’ diet choices, available ingredients, and kitchen appliances – offering a unique and personalized way of cooking. To streamline the cooking experience even further, tasteTAILOR seamlessly integrates its ingredient list with partner supermarkets. TasteTAILOR even offers a social community where users can share their experiences in the kitchen, get inspiration from other users, and connect with individuals with similar culinary interests.

Objectives and Value Proposition

TasteTAILOR’s most important objective is to grant customers the highest level of user satisfaction possible by making meal planning easy, seamless shopping through APIs, and engaging with the community through food recommendations powered by GenAI. Fundamentally, tasteTAILOR is committed to reducing food waste and fostering sustainable and healthy cooking with recipes to fit every taste.
We at tasteTAILOR present a multi-dimensional value proposition. We propose new dishes tailored to specific tastes but it will also generate shopping lists in supermarket applications in a hassle-free manner. Our customers can benefit from the step-by-step visual and audio instructions so that cooking can become enjoyable for everybody. The community feature provides an exciting way to connect with other like-minded culinary enthusiasts. 

Target Customer Segments

TasteTAILOR’s services are a perfect fit for several customer segments, for instance, busy professionals who have limited time to cook  but desire healthy recipes, students who are on an extremely tight budget and advanced hobby cooks who want to experiment with different types of cuisine and connect with other home cooks. The application also applies to people with special diet needs, making it the most suitable solution for those looking for a personalized cooking experience.

Key Activities and Resources

Key activities of tasteTAILOR include the design of customized recipes, using the power of GenAI technology. The personalization power comes from the analysis of user behavior, tracking one’s activity in the community, and through capturing user’s (daily) preferences.  Logically, the key resource of tasteTAILOR is the GenAI system that is woven throughout the entire business model, from generating recipes to chatbots intended for user interaction, and analytics that are inducted continuously for better personalization.

Challenges and Solutions

While the integration of GenAI offers significant benefits to the tasteTAILOR platform, it also comes with potential pitfalls, such as data privacy issues or technical glitches. The platform has taken stringent actions to consider these risks, implementing robust security measures and algorithm enhancement. The education of users through tutorials and community building is another crucial component in the adoption of AI-powered cooking solutions.

The Future of Meal Planning

Among the new generation of meal-planning applications and technological substitutions in a broader perspective, tasteTAILOR stands at the very top of elevating one’s cooking experience. TasteTAILOR offers personalized solutions that are not just tailored to the individual but are also ecologically-conscious. Taking everything into account, tasteTAILOR is not a meal planner; instead, it is a community-driven solution to empower its users to cook healthy recipes based on their own preferences. TasteTAILOR powered by GenAI is making cooking joyful.

Please rate this

Innovating Learning with Canv-AI: A GenAI Solution for Canvas LMS

17

October

2024

No ratings yet.

In today’s educational landscape, generative AI (GenAI) is reshaping how students and instructors interact with learning platforms. A promising example is Canv-AI, an AI-powered tool designed to integrate into the widely used Canvas Learning Management System (LMS). This tool aims to transform both student learning and faculty workload by leveraging advanced AI features to provide personalized, real-time support.

The integration of Canv-AI focuses on two primary groups: students and professors. For students, the key feature is a chatbot that can answer course-specific questions, provide personalized feedback, and generate practice quizzes or mock exams. These features are designed to enhance active learning, where students actively engage with course material, improving their understanding and retention. Instead of navigating dense course content alone, students have instant access to interactive support tailored to their learning needs.

Professors benefit from Canv-AI through a dashboard that tracks student performance and identifies areas where students struggle the most. This insight allows instructors to adjust their teaching strategies in real-time, offering targeted support without waiting for students to seek help. Additionally, the chatbot can help reduce the faculty workload by answering common questions about lecture notes or deadlines, allowing professors to focus more on core teaching tasks.

From a business perspective, Canv-AI aligns with Canvas’s existing subscription-based revenue model. It is offered as an add-on package, giving universities access to AI-driven tools for improving educational outcomes. The pricing strategy is competitive, with a projected $2,000 annual fee for universities already using Canvas. The integration also brings the potential for a significant return on investment, with an estimated 29.7% ROI after the first year. By attracting 15% of Canvas’s current university customers, Canv-AI is expected to generate over $700,000 in profit during its first year.

The technological backbone of Canv-AI relies on large language models (LLMs) and retrieval-augmented generation (RAG). These technologies allow the system to understand and respond to complex queries based on course materials, ensuring students receive relevant and accurate information. The system is designed to be scalable, using Amazon Web Services (AWS) to handle real-time AI interactions efficiently.

However, the integration of GenAI into educational systems does come with challenges. One concern is data security, especially the protection of student information. To address this, Canv-AI proposes the use of Role-Based Access Control (RBAC), ensuring that sensitive data is only accessible to authorized users. Another challenge is AI accuracy. To avoid misinformation, Canv-AI offers options for professors to review and customize the chatbot’s responses, ensuring alignment with course content.

In conclusion, Canv-AI offers a transformative solution for Canvas LMS by enhancing the learning experience for students and reducing the workload for professors. By integrating GenAI, Canvas can stay competitive in the educational technology market, delivering personalized, data-driven learning solutions. With the right safeguards in place, Canv-AI represents a promising step forward for digital education.

Authors: Team 50

John Albin Bergström (563470jb)

Oryna Malchenko (592143om)

Yasin Elkattan (593972yk)

Daniel Fejes (605931fd)

Please rate this

Learning how to code? Let Generative AI help you!

12

October

2024

No ratings yet.

When I started on my Python learning journey with Datacamp, I was excited, but I also faced challenges that tested my patience. As someone from a non-technical background, the structured logic of coding felt overwhelming. Even Python’s supposedly beginner-friendly syntax often appeared complex, especially when I encountered errors that I couldn’t quickly resolve.

Early on, one of the most frustrating issues I faced was debugging. It often led to roadblocks in simple mistakes like indentation errors or variable mismanagement. Despite the comprehensive learning modules on Datacamp, I often found it difficult to understand why my code wasn’t functioning as expected. Traditional resources usually provided solutions that didn’t quite align with my specific problem.

Moreover, applying theoretical concepts like loops, functions, and list comprehensions in practice was a significant challenge. While I could follow along with the lessons, I often found myself lost when it came time to solve problems independently. It became clear that I needed more personalized explanations to bridge the gap between theory and application.

That’s when I began using Datacamp’s integrated AI assistant, which proved to be a lifesaver. The AI provided on-demand explanations of the coding assignments, breaking down what each line of code was doing and helping me understand the purpose behind every function and operation. For example, when working on loops, the AI would offer examples and explain them in simpler terms, helping me grasp how to apply these concepts to real-world problems. It even helped me understand more complex concepts like recursion by providing step-by-step explanations and visualizations.

The AI didn’t just solve problems for me—it taught me how to approach coding challenges. Offering multiple ways to write a function or fix an error encouraged me to think critically about my coding style and improved my overall understanding.

I know this may sound like a promotion, but I genuinely recommend Datacamp to anyone interested in learning to code. It provides the most interactive learning experience, and the integrated AI makes the journey much smoother and more enjoyable.

Please rate this

Using Generative AI as My Marathon Training Coach

11

October

2024

No ratings yet.

I want to run a marathon next summer, so I turned to ChatGPT to help me out. Not with the running part, but mostly everything else. With generative AI, I can have a personal marathon coach that understands my needs, optimizes my training plan, and continually adapts as my fitness level evolves.

Setting Up My AI Coach

To create my AI coach, I used the MyGPT function offered by OpenAI. It allows you to customize a ChatGPT model to your liking by providing sets of instructions and information. I instructed it to act as my professional marathon coach who understands my goals, fitness level, and the specific challenges of marathon training. I explained my current abilities, body metrics, and general athletic history. I also provided my time goal, the marathon course info, and the date of the event. Moreover, I uploaded important data:

  • Running metrics for current, regular runs (distance, time, heart rate zone info, cadence, pace, split pace, effort level, etc.)
  • Apple Watch training data
  • Calendar data

This data ensures that the model can give more specific advice and training plans. For example, it uses my previous data to set pace targets considering terrain and what heart rate zone I should be in for a particular session.

Phased Training Plan

The AI-generated plan broke my preparation into phases, each with a specific focus. The initial phase was about building endurance with consistent mileage and easy runs. Later phases introduced speed work and tempo runs to improve stamina and pace. The AI explains the purpose behind each phase, which keeps me motivated and committed.

The phased approach means I’m not overwhelmed. Instead of seeing it as one long journey, I focus on each phase, trusting that the AI has mapped out the best way to build my fitness gradually. Each phase has unique challenges and goals, which helps me stay focused and structured. Here is an example of what it created:

Phase 2: Strength & Speed Building (Weeks 11–20)

Goal: Increase mileage, speed, and introduce hill work. Continue to improve cadence and manage heart rate.

Weekly Structure:

  • 5 days running:
    • 1 long run (gradually increasing distance)
    • 1 interval session (e.g., 4 x 1 km at 5:00/km)
    • 1 tempo run at marathon pace (5:40–5:50/km)
    • 2 easy runs (HR < 150 bpm)
  • 1–2 days cross-training or strength training focused on core and lower body strength.

Example Week 15:

  • Monday: Rest or cross-training
  • Tuesday: 8 km easy run (HR < 150 bpm)
  • Wednesday: Interval run: 5 x 1 km at 4:55–5:05/km with 2-min recovery
  • Thursday: Cross-training (strength-focused)
  • Friday: 8 km tempo run at marathon pace (5:40/km)
  • Saturday: 18 km long run (pace 6:10–6:30/km)
  • Sunday: Rest or mobility work

Long Run Progression:

  • Week 11: 18 km at 6:10/km
  • Week 14: 22 km at 6:00–6:20/km (HR under 160 bpm)
  • Week 17: 26 km at 6:10/km with last 3 km at marathon pace (5:40/km)
  • Week 20: 30 km long run (gradual build, HR < 160 bpm)

End of Phase Goal:

  • Cadence: Improve to around 165–170 SPM.
  • Long Runs: 30 km long run with a fast finish.
  • Tempo Pace: Consistent pacing at 5:40/km for up to 12 km.

How AI Adjusts Based on Real-Time Data

The AI adjusts based on real-time data from my Apple Watch, including cadence, heart rate, pace, and distance. If my heart rate is higher than usual during long runs, the AI might recommend more rest or adjust upcoming workouts. On strong days, it suggests more intensity or extra mileage. These adjustments keep my plan in sync with my body’s needs. The AI doesn’t just follow a set schedule—it adapts continuously to prevent burnout or injuries and to push me when I’m ready.

Personalized Insights & Accountability

The AI can provide insights into my performance, spotting patterns I might not notice. I can ask it to analyze my recent running data and give me advice if it is necessary to adjust the training plan. For example, when my pace stagnated, the AI suggested changes—more recovery days and different speed workouts. This helped me break through the plateau, improving my pace within weeks. These insights are invaluable and something I wouldn’t easily notice on my own.

Looking Ahead: My Marathon Journey Continues

With about eight months until the marathon, I’m excited to continue building on the progress I’ve made. Having a personalized plan gives me confidence that I will be well-prepared when race day arrives. Generative AI has transformed my training goals. It’s personalized, adaptive, and insightful. It helps me understand my capabilities, push limits, and gain confidence in my training approach.

Have you ever considered using AI to help with your fitness goals? I’d love to hear your thoughts or experiences—feel free to share in the comments below!

Please rate this

My Personal Chef

10

October

2024

5/5 (2)

Since its launch, I’ve mainly used ChatGPT to help me summarize academic articles and brainstorm ideas. Yesterday, I decided to let the LLM help me decide what to eat that night (OpenAI, 2024).

I rearranged the ingredients in my fridge so the main items were visible, moving condiments toward the back. I took a picture and uploaded it to ChatGPT with the prompt:

“What can I make for dinner tonight with these ingredients? I’m willing to go to the store for some additional ingredients needed.”

That’s it—just as if I asked a friend on the phone for a suggestion on what to make for dinner tonight, I asked the OpenAI model.

What it returned blew me away: (please note that formatted the output for visual appearances in this blog post).

“Here’s a simple recipe using the ingredients you have, plus some optional additions you could pick up from the store.

Butterscotch Pumpkin and Vegetable Curry

Ingredients from your pantry/fridge:

  • Butterscotch pumpkin
  • Carrots
  • Zucchini
  • Red bell peppers
  • Onions
  • Potatoes (optional, for a heartier curry)
  • Curry powder or paste (red or yellow)

Additions from the store (optional):

  • Coconut milk (for a creamier curry)
  • Fresh cilantro (for garnish)
  • Protein like chickpeas, tofu, or chicken

Instructions:

  1. Prepare the Veggies: Peel and cube the butterscotch pumpkin, and chop the zucchini, bell peppers, carrots, and potatoes.
  2. Sauté the Aromatics: In a large pot, sauté the onions in oil until they’re translucent.
  3. Add Spices: If you have curry powder or paste, add it to the onions and cook for a minute to release the flavours. If you’re using the peanut sauce as a curry base, add it now along with a dash of water.
  4. Add Veggies: Add the butterscotch pumpkin, carrots, and potatoes (if using). Let them cook for a few minutes.
  5. Simmer: Add coconut milk (if using) or water/stock, and simmer until the pumpkin and other veggies are tender (about 20 minutes).
  6. Add Other Veggies: Toss in zucchini and bell peppers, letting them cook for the last 5-7 minutes.
  7. Season: Taste and adjust seasoning with salt, pepper, or a touch of sugar if needed.
  8. Serve: Serve your curry over rice (you have poke rice) and garnish with fresh cilantro if you like.

    This will result in a rich, creamy, and slightly sweet curry with the pumpkin flavor shining through!Would you like to go ahead with this, or need adjustments?”

Amazing, right?! I was instantly inspired to cook this delicious-sounding meal and ran to the store for some chicken thighs and cilantro. Lucky for me, I already had the coconut milk, which the AI didn’t recognize in the picture.

For this blog post, I asked GPT to create an image inspired by this recipe:

While writing this blog post, I wondered if ChatGPT could give me an accurate estimate of the nutritional value of the dish, so I asked:

Caloric breakdown (per serving):

  • B. Pumpkin (100g)
    • Calories: 45
    • Carbohydrates: 11g
    • Protein: 1g
    • Fat: 0g
  • Carrots (50g)
    • Calories: 20
    • Carbohydrates: 4.5g
    • Protein: 0.5g
    • Fat: 0g
  • Zucchini (50g)
    • Calories: 8
    • Carbohydrates: 1.5g
    • Protein: 0.5g
    • Fat: 0g
  • Total:
  • Calories: 605 kcal
  • Carbohydrates: 86g
  • Protein: 28.8g
  • Fat: 15.5g
  • Bell Pepper (50g)
    • Calories: 15
    • Carbohydrates: 3g
    • Protein: 0.5g
    • Fat: 0g
  • Onion (30g)
    • Calories: 12
    • Carbohydrates: 3g
    • Protein: 0.3g
    • Fat: 0g
  • Coconut Milk (50ml)
    • Calories: 90
    • Carbohydrates: 2g
    • Protein: 1g
    • Fat: 9g
  • Curry Paste (15g)
    • Calories: 25
    • Carbohydrates: 5g
    • Protein: 0.5g
    • Fat: 0g
  • Rice (200g cooked)
    • Calories: 260
    • Carbohydrates: 56g
    • Protein: 5g
    • Fat: 0.5g
  • Chicken Thighs (100g)
    • Calories: 130
    • Carbohydrates: 0g
    • Protein: 19g
    • Fat: 6g

I am stunned! I checked the values online, and the caloric values only differed slightly (Grutto, 2024a; Voedingscentrum, 2024b). Not only does it give dinner suggestions and inspire me to make dinner with the ingredients I already have at home, but I can also keep a rough estimate of the macros I’ve consumed.
Imagine being on a diet and going out to eat with a friend. You want to keep track of your caloric intake but don’t want to bother bringing a scale to weigh your food with every new dish. Now you just take a picture, upload it to your Instagram story, and later upload it to ChatGPT to make a nutritional value estimate for you.

While this experience was undeniably impressive for me, it’s important to consider some limitations. For instance, the AI didn’t recognize the coconut milk in my picture, which was a key ingredient in the recipe it suggested. This highlights that image recognition technology isn’t foolproof, and you cannot rely on it 100%.
Also, while the nutritional estimates were close to official sources, they weren’t exact. For those with strict dietary requirements or allergies, relying solely on AI for nutritional information might not be a wise idea.
Lastly, uploading photos of your fridge or meals means sharing personal data with an AI service and thus can be a privacy concern. It’s important to be mindful of what you’re sharing and know how that data might be used or stored.

My experiment with using ChatGPT as a personal AI-chef was both enlightening and exciting. The ease of requesting a tailored dinner suggestion and a nutritional breakdown based on the contents of my fridge, shows me the potential of AI in everyday life. While there are limitations to consider, the benefits show a glimpse of the exciting future to come.

In the end, I find it amazing how the technology I first only use as a study and search tool, can also inspire me in other parts of everyday life. With my personal AI-chef, dinner dilemmas are a thing of the past.

Bibliography

Grutto, 2024, Bio Kipdijfilets bereiding en informatie Grutto! Available at: https://www.grutto.com/nl/vleesstuk/bio-kipdijfilet.

OpenAI, 2024, ChatGPT Available at: https://chatgpt.com.

Voedingscentrum, 2024, Hoeveel calorieën zitten erin? – Caloriechecker | Voedingscentrum Available at: https://www.voedingscentrum.nl/nl/service/vraag-en-antwoord/gezonde-voeding-en-voedingsstoffen/hoeveel-calorieen-zitten-erin-.aspx.

Please rate this

Why OpenAI’s Text-to-image Model Falls Short

10

October

2024

No ratings yet.

Since OpenAI launched its text-to-image GenAI tool, DALL-E, I have been using it for quite a while. Initially, I was amazed by how accurate and impressive the tool is for creating an image based on a short prompt, accurately depicting what I had expected in my mind. However, after using it a bit more, I started to realise that DALL-E has its limitations and is not yet ready to completely replace graphic designers or artists. 


If you have been using DALL-E for a while, you have probably experienced it creating weird texts in the image. As shown below in image 1, (prompt: Create an image of a tree in autumn in the Dutch city Utrecht, there are some stores with storefront names in the background), the store names are inaccurate, while the image itself looks amazing and of high quality.

Image 1: Created by DALL-E

Why can DALL-E create amazing visual images, but can’t produce normal text on these images while this seems so easy?

This is due to several technical issues. While the model is effective at understanding and generating visual elements based on prompts, it often lacks the capabilities to distinguish between visual content and written content. This limitation is rooted in the training technique of the models, when text is rendered as part of the image it becomes a visual pattern instead of a linguistic one. OpenAI has said that the next version, DALL-E 4, will have better results in distinguishing linguistic vs visual elements.

Another important issue to address, is the biased results from DALL-E. When asking to create an image of a CEO at work with an assistant, see image 2 (prompt: Create an immage of a CEO at work with an assistant), it will show a image of a man as a CEO and a woman as an assistant. This is because the model is trained on data from databases, where CEO’s are often represented as a man instead of a woman. Leading to biased and discriminatory results, ultimately reinforcing outdated gender stereotypes. The same counts for culturally insensitive and inappropriate results, because the model is not adapted to the cultural awareness of humans.

Image 2: Created by DALL-E

To conclude, these are just two of DALL-E’s limitations, while there are many more that I haven’t discussed. It is clear that DALL-E is not perfect and needs to improve their model before it can replace graphic designers or artists. Nonetheless, the potential for the future is immense, and for now, it is incredibly fun to experiment with.

Thanks for reading, if you have an opinion about this topic, please leave a comment below.

(Disclaimer: this blog is written based on my personal experience)

Please rate this

AI Meets the Kitchen

10

October

2024

No ratings yet.

When I was younger, I used to dislike cooking. It always seemed overwhelming to me since there were so many different techniques, recipes and ingredients. I was often confused, especially when trying to come up with meals based on whatever I had at home. However, when I began exploring different generative AI tools, my perspective shifted completely. These AI tools transformed cooking into something enjoyable and creative rather than a chore I needed to finish as soon as possible.

One of the most valuable tools in my cooking journey has been ChatGPT. Instead of spending hours browsing cooking websites or watching YouTube videos, I can simply ask ChatGPT what to cook using the ingredients I have at home. It generates personalized recipe ideas along with step-by-step instructions, which saves me time and reduces food waste by helping me make the most of what’s on hand. Additionally, I use it to create customized meal plans tailored to my dietary goals, such as increasing protein intake or following a gluten-free diet. Below is an example of a weekly meal plan I’ve created with ChatGPT’s help.


I also believe that generative AI can be a useful tool for menu and recipe creation in restaurants. By generating novel recipes based on specific criteria like ingredient availability or dietary preferences, AI helps restaurants stay competitive and offer unique menu options. It can assist chefs by suggesting adjustments, reducing the time needed for recipe development. AI simplifies parts of the creative process, allowing restaurants to consistently introduce new dishes while reducing the pressure on the culinary teams.

As generative AI continues to evolve, its potential to revolutionize at-home cooking and the restaurant industry continues to grow. Whether you are a beginner like myself or a professional chef, I believe that generative AI can enhance everyone’s cooking experience.

Please rate this

AI and the automation of accounting 

9

October

2024

No ratings yet.

AI has changed the landscape in many work fields, such as in data, law, finance and even accounting. Recently I have taken an assistant-accounting job at a firm paired with my master studies. I have discovered a lot about AI and what it could do for accounting that I thought were not possible.  Having been there for 2 months, I have learned a lot about accounting itself, but also what AI did for me while I was at the job. According to an article by the open journal of business and management, accountants become more efficient in handling the bookkeeping and become more productive in their work due to the use of AI (peng et al, 2023). I found this to be true as well. 

The first week I began working the invoices in the bookkeeping system, I was very reliant on my own knowledge. I tried categorizing each invoice in a specific item for the ledger, which was confusing sometimes. Because there were a lot of different items and different kind of invoices. With the help of ai however, it showed me what type of invoice belongs to what items in the ledger.  For example, one particular invoice showed me a bill of the restaurant where a particular client of the accounting company was having dinner. At first, I was not sure where this type of cost could be booked in the ledger. I asked AI: “at which item does a restaurant bill belong to in a ledger?” The prompt I gave ChatGPT provided me with this answer:

I asked my employer, who is an RA (registered accountant) for verification is this was true, he confirmed that it was indeed true. From then on, I started using AI a lot more for bookkeeping. Especially for items I was unsure of. It helped me become more knowledgeable, but also helped me to become more efficient and productive. 

Even though there are a lot of benefits of using ai in accounting, there are also downsides. Each bookkeeping for a specific client is different. Tailoring to these clients cannot be done always done through AI. My employer explained to me that the accuracy of knowledge AI has on accounting and bookkeeping is broad, but that I should not always rely on it to keep the books. Because clients sometimes require specific wishes in what purpose their ledger serves and on what books costs are categorized. For example, some clients want to receive a larger tax return. Therefore, they would categorize some cost to a specific item in the ledger that are relevant for that outcome, and others use It to justify cost that they made for the past few months that could be lost in other items, due to inaccuracy of the AI. This type of variety is sometimes very confusing for the AI making some prompts not always accurate to what the client needs. As accepting is also a type of consulting and advisory to companies.

In my opinion, AI is beneficial for the future of financial bookkeeping, and it will probably change a lot of aspects in the financial field. I do however think that when it comes to personalized tailoring to clients with jobs such as consulting and advice, especially in the tax field. It is still of relevancy that accountants or financial advisors take responsibility in helping clients themselves to keep them satisfied and use AI to their benefit. Becoming more efficient and accurate themselves due to the clever use of it.

Peng, Y., Ahmad, S. F., Ahmad, A. Y. B., Al Shaikh, M. S., Daoud, M. K., & Alhamdi, F. M. H. (2023). Riding the waves of artificial intelligence in advancing accounting and its implications for sustainable development goals. Sustainability, 15(19), 14165

Please rate this

My Experience with GenAI: Improving Efficiency or Becoming Stupid?

9

October

2024

No ratings yet.

I work as a part-time data analyst at a software company, where I analyze sales data. My 9-5 mainly consists of writing code, specifically using SQL in Google Bigquery and creating dashboards in PowerBI. I love using GenAI to help me write queries faster which would have taken me a long time to compose by myself. Additionally, I am a student and use GenAI to help me better understand course content or inspire me on what to write about during assignments. Generally, I would say that GenAI benefits my life as I can get more done in less time, however, from time to time I start to question whether I am not just becoming lazy.

I use GenAI on a daily (almost hourly) basis and rely on it in many ways. I mainly use ChatGPT 3.5, when ChatGPT 4o’s free limit has been reached, and Gemini, when ChatGPT is down. Based on my own experience, I can say that being good at ‘AI prompting’ is a real skill in the field of data analytics as it can drastically improve the efficiency with which you write queries, and therefore, the speed with which you finish tasks. My manager recently even held a knowledge-sharing meeting in which he discussed the best practices to use for data analysts when interacting with ChatGPT. Using GenAI has become a real thing in the field of data analytics, and is not something to be ashamed of.

However, I cannot help but sometimes be slightly embarrassed when I read back the questions I’ve asked ChatGPT. It seems that with any task that requires a little bit of effort or critical thinking, I automatically open the ChatGPT tab in my browser to help me come up with the right approach to solve the task at hand. I don’t even try to solve things by myself anymore, which makes me question: is this something to be desired?

The image presents an interaction with ChatGPT regarding the risk of using GenAI on human intelligence.
The image presents an interaction with ChatGPT regarding the risk of using GenAI on human intelligence.

As explained by ChatGPT in the image, using GenAI indeed frees up more brain space for things that are important. If I can use less time to get more work done, this improves my work efficiency and also gives me more time for things that I find more valuable, such as spending time with family or friends. Right now, it is still too soon to be able to determine the impact that using GenAI will have on our own (human) intelligence. In the meantime, we should just continue using it for repetitive tasks that would normally take much of our valuable time and hope that it is not ChatGPT’s plan to stupidify humanity before it can take over the world.

Please rate this