Your Guide to AI
If you’re reading this I’m positive you want to learn about Artificial Intelligence (AI) and if I may be so bold you are also concerned about what may happen to you in the future. If so please be sure to read this guide in its entirety as I will cover what AI is, the history of AI, how you can use AI to get a head start, the different AI available right now for you to use, the future of AI as well as actionable steps you can take to prepare yourself.
Let’s get started.
What is AI
In Layman’s terms, Artificial Intelligence (AI) are machine algorithms that can simulate human intelligence. AI can learn from repeating iterations of given tasks. While this sounds incredibly simple the technology behind it is far from it.
Different types of Artificial Intelligence
AI can be characterised based on complexity and goals.
Artificial Narrow Intelligence (ANI)
Also known as Narrow AI, this type of AI is made to achieve very specific goals. A great example is face recognition. AI is used here to help match facial characteristics with an original image, the AI used for facial recognition is tasked solely to recognise faces and no more. Giving an AI a specific goal and boundaries is what makes it a Narrow AI.
Artificial General Intelligence (AGI)
AGI also known as Deep AI, this is a concept where AI can imitate human intelligence. Unlike Narrow AI, Artificial General Intelligence (AGI) will have limited boundaries and no specific goal in mind, it’ll be the definition of a jack of all trades.
Researchers have yet to achieve this level of AI but once they do it will cause more disruption than the likes of Chat GPT and similar AI.
Large Language Model (LLM)
Large Language Models are programs that have been fed huge amounts of data (often from the internet) as such can recognise and interpret human language. The more curated this type of AI is the more accurate the responses will be, the broader the less accurate.
History of AI
AI didn’t just appear from thin air, in the 1950s researchers started research on Deep AI (AI capable of generally performing any intellectual task). The lack of results in creating Deep AI eventually led to the creation of Narrow AI (AI built for a specific purpose). Until 1980 research was split between these two fields.
Around 1980 “Machine Learning” became the key area of research giving computers the ability to learn. Machine Learning is a subset of AI, while it does not replicate human-like intelligence machine learning gives computers the ability to learn something based on a specific goal and set parameters.
Without getting into the science of it, in the early 2000s researchers created new methods that fundamentally changed Machine Learning which helped create the systems we know today; facial recognition, identifying tuberculosis (with 96% accuracy), self-driving cars, and more.
In 2022 Open AI released Chat GPT a Large Language Model or LLM. The release of this AI shook the world and has caused a huge wave of competing technologies (Google Bard and so on) which prompted companies to incorporate AI within their business functions. Which brings us to today where we now have 100s of AI to choose from, all with their own specialisations.
The future of AI
So what should we expect from the future of AI? There are two things to keep an eye on and you may be able to guess. Much like in the 1950s Deep AI or AGI will be a massive focus for researchers going forward, and in the future when it is released Artificial General Intelligence (AGI) will cause even more disruption than Chat GPT in 2022. Making it more important that you gain an understanding of this technology.
The other area to focus is Artificial Narrow Intelligence (ANI). ANI or Narrow AI is being implemented into almost everything, from video creation, image generation, content writing, CAD, Website building and more. It feels like new ANI models are coming out almost weekly now.
Will AI take your job?
Now let’s answer the big question that everyone is worried about, will AI take your job? Yes and no. In its current state, many AI tools require human input to be able to generate anything and what is generated can be far from perfect, however, saying that there is no denying AI’s ability to compute data and to spitball ideas.
In the future (I don’t know when exactly, I’m not a time traveller), AI will reach a point where it is capable of performing tasks better, faster, and autonomously without human input. Once at this point, AI will have all the capabilities to take over many jobs currently occupied by people.
Of course, it is sad and upsetting that people may one day be replaced by machines however, this does not have to become a reason for chaos. In a dystopian future where most if not all tasks were completed by AI that would give us more time for other activities like; spending time with family and other passions.
When I am talking about the topic of world altering technologies I like to give the following example; 150 years ago horse and carriage were the main transport for people then all of a sudden in 1886 the first modern car that was practical and marketable was invented, and soon after roads were taken up by mechanical boxes on wheels instead of fluffy horses trotting down a cobbly street. With the decline in horses, there was less need for stables, stable keepers, and horseshoe makers. Thousands of people around the world lost their jobs and had to go into a different industry.
This type of progression has happened throughout the entirety of human history with old industries being taken over by newer and better ones. Every time that happens people always find a way to adapt to the new environment, this is one of our greatest abilities as humans, the ability to adapt to changing environments.
Going back to the main point, in the near term AI will not take your job however, it will more than likely be implemented into your role as a way to make your work more efficient. In the long term as AI develops it will reach a point where is it capable of performing tasks without human input, once at this point we will start to see jobs being overtaken by AI. It is for this event that you need to be equipped with the knowledge to become irreplaceable by AI.
How to position yourself to help navigate the future?
Research, optimism and effort
You need to research AI, what it is and how it will affect your industry. When finding this knowledge you need to view it with an optimistic bias, what ways can AI help you or your industry improve, and what do you need to learn to help enforce that improvement? Finally, you have to put effort into learning and implementing these changes, simply leaving this up to other people will put you in a vulnerable position of being replaced by someone who has your knowledge on top of AI.
How to use AI
Now that you know more about what AI is, where it came from as well as future implications of such a technology we will now start learning how to use AI.
Let’s begin with the well-known ChatGPT
ChatGPT operates off of a simple Prompt-and-response system. You write a prompt, which can be a question or a task, and then the AI will generate a response based on the context of the prompt and its own learned data.
This sounds super simple, that is because it is. The skill lies in your ability to create an accurate prompt so that the AI can generate the best response/answer to your prompt/question.
The Art of Prompts
Example 1; “Can you write a blog about AI”
ChatGPT will then go away and generate a very generic blog talking about AI. Since you have not given it much context the response generated will lack lustre.
Example 2; “Can you write a blog about how AI will affect those working within internal communication roles”
This prompt is much better than the first as it offers more context about the target audience for the piece as well as a more specific topic.
The amazing thing about AI is that it is capable of making adjustments to previous responses as long as you ask the right prompt.
Example 2a; “Write a paragraph at the end of the blog with compelling call-to-action messages that lead readers to my landing page about [insert landing page details]”
This follow-up prompt is the most advanced out of the three prompts given thus far, not only have we asked ChatGPT to generate a new paragraph at the end of the previously generated blog but we have asked for it to be specific, “compelling call-to-action messages that lead readers to my landing page”. We could have left that prompt there however, the response generated by ChatGPT would have been too generic and lacked context which I would need to add later on.
Adding that last bit of the prompt gives ChatGPT the context required to generate an accurate response to my wants.
ChatGPT Limitations
Every AI has its limitations ChatGPT is no different
- There is a possibility for misinformation
- It can generate inaccurate information
- ChatGPT 3.5 (Free Version) only has information dating up to 2021
AI is not perfect and still requires human input to generate responses and to ensure the responses are accurate.
This brings up a golden rule of AI, at least for AI as it is now, don’t take AI’s responses as gospel. Always read through the response and ensure that it is accurate.
Other AI
So far we have only been talking about ChatGPT, the skills you learn here with prompt creation will carry over to other similar AI models, Google’s Bard and Microsoft’s Bing (based on ChatGPT).
However, AI is not limited to text-based responses. DALL.E 2, MidJourney and Leonardo AI are capable of generating images based on text input.
Even commonly used platforms like Canva, and selected Adobe products have incorporated AI into their software.
Canva
Canva is many people’s preferred design platform as it is easy to use and offers a huge library of templates and other designs to kickstart projects. More recently Canva has added an AI image generator as well as a video generator.
While the output has changed from text to image and video, you still use prompt-based systems to generate these images/videos. However, conversely to the prompt we used for ChatGPT, AI which generate images/videos require more descriptive words, the better you are at painting a scene with your words the more accurate the image/video will be.
Example 1: “Candle flickering”
Example 2: “Crowded shopping district in London with clothing shops in the background”
Key takeaways and actionable steps
There are hundreds of AI out there and more and more businesses are incorporating AI into their products and operations. While we mainly covered AI which can be used externally or for internal communications there are AI that are purpose-built to help with business operations and more technical roles such as coding.
The first thing you need to do is find out which AI you can use to enhance your work.
Secondly, you need to use those AI software wherever you can to better learn the intricacies of them.
After you’ve gained some experience using AI and have improved your prompt engineering skills, you need to think about how AI can be used to supplement your role. Not take over your role, but to supplement certain aspects of it. For example; maybe there are certain data-crunching aspects of your role where you need to manually create a manager report, can you use AI to automatically generate manager reports every month giving you more time to do other tasks?
After research comes implementation. Have discussions with leadership presenting your findings and offering suggestions of where AI can be implemented to improve your role. Remember AI is a tool that can be used to enhance your craft.
To get you started on your journey to master AI here are some key resources.
AI Resources
How internal communicators can use AI – How ChatGPT Will Revolutionise Internal Communications – NewZapp Communications
Learn how to engineer prompts w/ Google – Prompt Engineering for Generative AI | Machine Learning | Google for Developers
Text generation: Chat GPT – https://chat.openai.com/
Image generation: DALL.E 2 – DALL·E (openai.com)