AI in the Workplace. A Short Guide For Beginners on How to Start Using Artificial Intelligence in Your Daily Work.
With the sudden surge in the popularity of AI-powered tools, it might seem like everyone around you has become an instant expert in the field. If you feel like you're lagging behind, this brief guide on integrating artificial intelligence into your daily work is just what you need. Let's explore practical applications of AI technology in the workplace and beginner-friendly tips to enhance your skills and boost productivity.
This article is part of the new Decoding AI series by hero/dot, which focuses on showcasing the practical aspects of Artificial Intelligence and its impact on the future of business and IT. Decoding AI stems from our commitment to keeping up with the latest technological advancements. In monthly internal workshops, our team members share their experiences with different aspects of using AI in their daily work. We then condense that knowledge into helpful articles to share with you on our blog.
If you require a more technical introduction to artificial intelligence, we recommend reading our dedicated article on the topic. Here, our focus will be exclusively on the practical applications of this innovative technology. Be sure to read until the end to:
- Discover real-life examples of using AI tools for various work-related tasks.
- Level up your ChatGPT skills with insights on prompting, plugins, and the temperature parameter.
- Familiarize yourself with basic safety guidelines for using AI in the workplace.
- Find valuable resources to expand your knowledge.
But First, Change Your Approach to Artificial Intelligence (AI)
Artificial intelligence (AI) has been a trending topic for some time now, particularly in discussions about the future of work. The dust hasn't settled yet, as extreme pessimists battle extreme optimists. On one side, we find visionaries who believe that AI will usher humanity into a prosperous and productive golden era. On the other side, many workers feel like the widespread adoption of AI will result in significant job losses and a global recession.
We want to assure you that, at this moment, AI is not a direct threat to your job. However, those who have learned how to use it to their advantage do pose a challenge. Instead of dreading this change, let's embrace it and acquire new skills to remain competitive in the ever-evolving job market.
How Can I Use AI in The Workplace?
Do you ever find yourself wasting precious time on mundane, repetitive tasks rather than focusing on projects that truly require your expertise? Using AI to handle tedious or monotonous assignments is a smart hack for increased productivity during work hours.
- Writing E-mails: Generate e-mails or craft responses within minutes. Simply provide a rough outline of the information you wish to convey and select an appropriate tone. You can even experiment with AI to add humor, professionalism, or a touch of passive-aggression to your messages.
- Recording Meetings: The days of frantic note-taking during meetings are a thing of the past. For instance, tl;dv can convert recordings into transcriptions, summaries, and agendas, while also organizing them by topic or time.
- Creating Summaries: There are plenty of AI-power tools out there that can condense lengthy YouTube videos into short summaries or extract the key points from written reports. Artificial intelligence can assist you in making better decisions about which text or video resources are worth your attention.
- Translation: AI chatbots offer an advantage over traditional translators as they excel at grasping the context behind the text. You can even conduct a simple experiment by comparing the results of running the same message through different translation tools.
Whether you're dealing with creative block or seeking to streamline your workflow, integrating AI into your creative toolkit offers numerous benefits. You might be surprised by the ingenious results that can emerge when human intelligence and artificial intelligence collaborate.
- Seeking Inspiration: When you're feeling stuck and traditional idea-generation methods aren't yielding results, AI can lend a helping hand. Ask a chatbot to generate concepts for illustrations, or provide a list of 10 title ideas for an upcoming blogpost.
- Writing Blog Articles: Using AI for crafting longer-form content isn't as straightforward as requesting ChatGPT to generate the entire thing. Fortunately, we've already published an article on this exact topic – you can check it out here!
- Finding Emojis: Sorting through hundreds of emojis to find the perfect one is a simple but time-consuming task. Next time, copy and paste your post to a chatbot and let it choose the most suitable emojis for you.
- Image Generating: From Midjourney to new features introduced by Adobe, technology consistently pushes the boundaries of digital graphic design. Now you can remove unwanted elements from photos, expand images beyond their frame, generate custom vectors, and more.
Data analysis is one example of AI's pivotal role in increasing productivity. By rapidly sifting through vast datasets and automatically identifying patterns, artificial intelligence helps save resources on time-consuming tasks.
- Data Analysis: Even ChatGPT can be useful in the analysis process, thanks to the Advanced Data Analysis plugin. However, for more advanced users, we recommend Power BI, which has recently added an AI-powered copilot function, as an excellent tool for examining, visualizing, and sharing business data.
- Knowledge Base: Imagine a knowledge base that can promptly address your queries on any company-related topic. Increasingly, organizations are employing AI to construct such databases, and this task will soon become even more accessible with Google's Bard.
If you want more tips and tricks on how to use AI in this area of work, be sure to follow our Business Analyst Filip Niedziela on LinkedIn for regularly posted valuable insights.
Artificial intelligence and software development go hand in hand, undoubtedly. While we plan to dedicate an entire blog post to this topic in the future, here's a brief introduction to the possibilities of AI in the realm of code assistance.
- Generating Code: Tools like GitHub Copilot employ AI to streamline development and enhance efficiency. They achieve this by suggesting code snippets, automating repetitive tasks, and offering increasingly accurate recommendations.
- Explaining Code: Good news for anyone who has struggled to understand someone else's (or even their own) code. AI-driven tools can now analyze and provide explanations for complex code structures, facilitating collaboration among developers.
- Fixing Bugs: AI can automatically detect and correct bugs or errors in code, elevating its quality and reducing the time and effort required for debugging and maintenance.
- Improving Code Efficiency: Receive suggestions for optimizing existing code, such as identifying redundant lines or proposing more efficient algorithms.
- Translating: AI can aid in the automatic translation of code between different programming languages. This allows you to reuse code across various platforms and migrate projects to different technology stacks more efficiently than ever.
Now that we've covered the basics, let's explore three ways anyone can advance from a beginner to an intermediate ChatGPT user: prompting, plugins, and temperature.
Get More Accurate Results with Prompting
Have you ever felt let down by ChatGPT's response and abandoned your quest for answers? Perhaps you even assumed that your request was too challenging for the generative AI to handle. Sorry to burst your bubble, but the issue likely doesn't lie in ChatGPT's capabilities, but in the instructions you provided. Fortunately, we have a one-word solution to this common problem: prompting.
For instance, let's say you need to compose a social media post about software estimation techniques. Here's what a beginner might type:
Write a social media post about software estimation techniques.
While the content of Chat's response to this request may suffice for your basic information needs, the format leaves much room for improvement. This is because, when aiming for highly specific results that align with your company's tone of voice and resemble previously posted content, you should craft more detailed prompts. Let's try again:
You're a copywriter responsible for producing content for a software house. Craft a 200-250 word post for LinkedIn discussing software estimation techniques. Be sure to:
- Select and describe the three most reliable estimation techniques, highlighting their advantages and disadvantages.
- Provide guidance on choosing the best technique for a specific project.
- Write in clear and concise sentences.
- Avoid overly formal language or complex vocabulary.
- Maintain short paragraph lengths and use bullet points for improved readability.
The second prompt surpasses the first by specifying the style of expression and the approach to the topic. Also, assigning a specific role to ChatGPT (in this case, the copywriter) is a valuable technique. Keep in mind that there's no one-size-fits-all method for expert prompting. We encourage you to experiment with various queries for the same task and see what works best for you. In the long term, the experience you gained will enhance your productivity, allowing you to either reuse certain prompts or create new ones more quickly. Moreover, we plan to dedicate an entire workshop to this topic in the near future.
Unlock New Possibilities with ChatGPT Plugins
Now that we've learned how precise prompting can improve the quality of AI-generated writing, can we enhance its mathematical abilities? Currently, when you ask ChatGPT to multiply 2433 by 3434, it incorrectly insists that the answer is 8,363,622. This is a classic example of what's known as ChatGPT hallucinations. This phenomenon occurs because equations involving large numbers are less common in its training data, leading to a lack of knowledge about the correct result, even if it can't admit it.
However, there's a simple trick to help artificial intelligence calculate such queries with pinpoint accuracy: plugins. In this case, we'll use a plugin called Wolfram Alpha. Running this plugin finally provides us with the correct result:
Lucky for you, the power of plugins doesn't stop at improving Chat's mathematical skills. Here are a few recommendations for plugins that can prove highly useful in your work:
- Access PDF & Docs allows ChatGPT to access PDF files on our Google Drive and respond to our questions with information from within these documents. It's a fantastic way to increase efficiency when collecting data from lengthy reports that would take forever to read. With Access PDF & Docs, you can simply ask artificial intelligence to scan the PDF and provide a summary.
- Make a Sheet enables you to generate a CSV file from Chat's responses and import it directly into Google Sheets or MS Excel, which can be extremely valuable for data analysis tasks.
- Diagrams: Show me does a similar thing but with exportable diagrams, charts, flowcharts, and mind maps.
- ScholarAI ensures that the answers to your questions are backed by scientific research by connecting artificial intelligence to a vast database of peer-reviewed research articles.
- Wikipedia is exactly what it sounds like: it allows users to search for general knowledge and current events information directly from Wikipedia.
There are countless other plugins available, and we encourage you to explore them yourself at the OpenAI Plugin Marketplace. You might even discover something useful for your personal life, like Kayak. Who would have imagined that artificial intelligence would one day reduce the time needed to plan a vacation, book a hotel, and find a flight deal!
Unleash Creativity with the Temperature Parameter
This tip may seem a bit advanced, but we believe it's helpful even for total beginners. ChatGPT features a modifiable parameter known as temperature, which influences the creativity of its responses. The temperature value ranges from zero to one. When set to zero, ChatGPT selects words and phrases with the highest probability of appearing in a specific order. While this doesn't guarantee identical answers to the same question every time, the neural network may take the response in a slightly different direction on each occasion. However, with a temperature of zero, the results tend to be rather generic.
As you might have already guessed, higher temperature values lead to more creative responses from ChatGPT. However, it’s important to remember that by raising the temperature we also increase the risk of hallucinations. To adjust the temperature, simply enter the command: "Set the temperature to X," and then proceed with your prompt. We recommend playing with this parameter to understand how artificial intelligence behaves at different temperature levels. Here's what you can expect from responses within these temperature ranges:
- 0 to 0.3: Lower temperatures yield more direct, focused, and coherent outputs.
- 0.3 to 0.7: Use medium temperature for balanced results.
- 0.7 to 1: High temperature values are ideal for generating highly creative and original, although potentially less factual, responses.
Take a look below to see the striking differences in responses to the same prompt at vastly different temperature levels:
Take a look below to see the striking differences in responses to the same prompt at vastly different temperature levels:
Prompt: Set the temperature to 0 and write four sentences about a dog.
Response: The dog is a domesticated mammal known for its loyalty and companionship to humans. It is a member of the Canidae family, often bred for various purposes, including hunting, herding, and as pets. Dogs come in a wide range of breeds, each with unique characteristics and traits, from the tiny Chihuahua to the majestic Great Dane. They have been our trusted companions for thousands of years, providing love, protection, and unwavering devotion to their owners.
Prompt: Set the temperature to 1 and write four sentences about a dog.
Response: The dog, a loyal and affectionate companion, bounded across the grassy meadow, its tail wagging with unbridled enthusiasm. With soft fur that seemed to glisten in the sunlight, this four-legged friend exuded an irresistible charm. Its expressive eyes sparkled with a mixture of curiosity and joy, a testament to its boundless energy and unwavering love for its human family. As it playfully chased a fluttering butterfly, the dog’s zest for life was a reminder of the simple and pure joys that our furry friends bring to our lives.
Don't Forget the Safety Rules
Rule #1 – Do Your Own Research
When ChatGPT first came into the spotlight, we all became fascinated by the remarkable results it could produce in a matter of seconds. However, let's not confuse probable answers with factual ones. It all boils down to how ChatGPT operates. In simple terms, as a language model, it utilizes a deep neural network to statistically predict the probability of a particular word appearing next in a given context. The reason Chat's responses are generally fact-based and human-like is that it relies on an incredibly vast dataset of various internet texts.
Lately, there's been much discussion about the so-called "hallucinations" of artificial intelligence, which refer to false statements without backing in the training data. Therefore, it's vital to DYOR (Do Your Own Research) and always fact-check the information provided by the chatbot. Asking ChatGPT can serve as a useful starting point for further research, as it offers you an outline of topics worth exploring in depth.
Rule #2 – Never Reveal Sensitive Data
There are methods for automatically masking sensitive data, but that's a topic for another article. For now, we recommend choosing an offline chat room, such as Faraday.dev, when deciding on which AI tools to use in the workplace.
Rule #3 – Follow Safety Guidelines
Last but not least, to further explore the topics we've covered in this article, here are some valuable resources:
AI For Anyone
Ai for Anyone is a 6-hour long online course created by DeepLearning.AI. It is designed for non-technical business professionals and aims to explain what AI is, its current capabilities and limitations, and how to apply it in a corporate context. We highly recommend it!
ChatGPT Explained: A Normie's Guide To How It Works is an article written by John Stokes. What sets this publication apart from others on the same topic is Stokes' clear and comprehensive (but not overly simplistic) presentation of the core concepts behind modern AI-powered chatbots.
In What Is ChatGPT Doing … and Why Does It Work? Stephen Wolfram takes a more mathematical approach to explain the inner workings of ChatGPT, complete with examples and diagrams. It's a lengthy read but a very insightful one.
While we plan to write a comprehensive article on prompting in the near future, here are some knowledge bases that cover the topic:
- learnprompting.org: An introductory document that covers various prompting scenarios with real-life examples. It's beginner-friendly and easy to understand.
- Prompt Engineering Guide: A more advanced but, nevertheless, a comprehensive guide that will certainly enhance your prompting skills.
- OpenAI: It offers 30 examples of possible applications of ChatGPT, complete with recommended queries.
HelloMidjourney is a free online course in Polish developed by Artur Kurasiński. If you're interested in creating AI-generated images with Midjourney, this is an excellent starting point for your creative adventure.
Stay Tuned for More Decoding AI!
If you found this first instalment of the Decoding AI series enjoyable, be sure to stay tuned for our upcoming articles. Every month, we will continue to offer a large amount of practical insights based on our team's expertise in the field. Furthermore, we release new AI-related posts, infographics, knowledge quizzes, and more on our social media platforms each week, so make sure to follow us there!
Need expert assistance with your digital project?