Writing Blog Posts Using Artificial Intelligence (Ethically!)

1. Introduction: Hi there, my name is Rebecca and I am new to teach you how to use AI reading tools for your blog. We're starting to see AI tools find a place in pretty much any industry, especially those that use some sort of technology or machinery. And as a writer, a blogger, or a market or anything like that, you are definitely already using different technologically advanced tools to do your work. Adding AI to the mix can be a great way to conquer writer's block, get more written faster, and generate new ideas that you may have never come up with otherwise. However, using an AI writing tool isn't as simple as pressing, generate and just using whatever comes out. Of course he could do that, but the text that you generate might not be that good. While AI tools are getting more advanced all the time, it's still can't exactly go writing unsupervised yet. When it comes to your blog, you have a point to make and you want to convey it correct information. And this is why it's important to use AI not as an authority on writing, but rather as a complementary tool to your writing process. In this course, I'm going to introduce you to some background information about AI writing tools and how they work. As we'll introduce you to a few different options for tools that you could try using yourself. But that foundation in mind, we're going to look at how to outline and structure a blog post in a way that makes it really easy to work with these AI tools. Then we're going to generate a fake blog post together using just some prompts that I came up with and go over some ways to use the writing tools in a way that makes them really effective. And also make sure that you are writing a high-quality posts for your blog. By the end of this course, you'll be able to use AI reading tools to get more written faster for your blog or website. 2. How AI Writing Systems Work: Let's get started by discussing briefly how AI writing tools work. These tools start with a collection of text or a corpus or a neural network. Those are some different terms for it that acts as a reference points for all the calculations and processes that follow different tools. We use different corpus databases taken from different places. One of the most popular projects out there is the GPT-2 and GPT-3 models. The latter being the most recent version, which is run by a company called Open AI or a project called Open AI. Gpt-3 is an autoregressive language model that uses learning to produce human-like texts. It's trained on 10 percent of the Internet. Like all the texts sits on the Internet, meaning it has an unthinkably huge amount of source material to generate from. So what this means is that the model will analyze the billions of words and documents that it has stored in its corpus to study patterns in language that allow it to write intelligently on a subject. If you give this model a prompt like a sentence starter or something like that, then it will figure out something plausible to come after those words and start writing for you using probability and some mathematics. So far a really simple example, Let's say that you're starting to write a sentence about a cat. And you want to prompt the model that you're working with to continue talking about that cap. So if you start the sentence, if you go and any number of directions to talking about different things related to cap. So what cats like to eat or how cats are different than dogs or famous cats at history, which may or may not be fictional, it can often create fictional names or events for people, but it's not likely to start saying something about cats that live on the planet Mercury, for example, because I suspect there isn't anything in its corpus about cats living on Mercury for obvious reasons. The exception could be, of course, if the corpus science fiction novels or any sort of material like that, anything, Cy5, cats, I guess. But even in this case it comes out to probability. So as you can see, there is actually a lot of math and probability that goes behind the scenes in any air reading tool. But luckily as writers, we don't have to worry so much and hope that part. There are plenty of tools that you can start using right away as a writer, the options will vary depending on what sort of task you want to perform and what language model process you want to use. So right now, I'm favoring a tool called intricate. This tool allows you to test it for free if you want to try that one, it gives you 40,003 characters before you have to subscribe. And for Kant uses the magnetron 11 bead neural network. Another tool is copy.txt II, which is specifically targeted at copywriters and marketers. So that might be more interesting if you are in that area. They use the GPT-3 model as their basis and they offer a free trial. And the other one is AI dungeon. This is a very popular with fiction writers, especially as it's set up sort of like a, not sort of like exactly like a text-based video game. And you can feed any sort of prompt into and see what you get. So this tool is actually free to use and operates on various tiers of the GPT-3 model with some possible subscription options. There are also many other tools out those is just a various, a sample of them and you can find them often by searching online for AI writing tools. So you can Google that. And this can generate lots of useful results and programs and more are being built all the time off of those different engines that we've talked about, especially for different niches of writing. Now before we get into how to work with these tools effectively for blog posts, Let's discuss the ethics, copyright, and use cases of the handwriting tools in the next video. 3. Ethics and Use Cases: The discussion around AI ethics is constantly evolving so much so that it's almost difficult to have a static conversation about it. But let's touch on a few issues so that you can be more aware of the changing landscape of AI as you start to use its tools. One important discussion around AI is in regards to copyright. Right now there's no clear ruling on if, how or when you would need to disclose that a text is written using an AI writing tool. Some people think that because the tools use and that include texts written by other people, that there should be disclosure. Other people argue that because these tools generate almost 100% plagiarism free text, and the fact that it is impossible to trace back something in a I wrote to a specific document. There's no need to worry about issues like copyright. If this text was copyrightable, then who would the copyright belong to? The person who's generating it or the person who are the team who builds a tool? Or would it be the people who's writing from across all the internet went to create the corpus. You can see the difficulty already. These are questions that people are going to be asking more and more MCC, AI interwoven into different facets of business and life. But for now there is no obligation whatsoever to disclose when or where AI is being used in your writing. In fact, you'd be surprised at how many news outlets and writers of their already co-writing with it. When AI writing tools first started to hit the market, there was a lot of concern that people would basically use these tools to keyword spam websites and to try and rank on Google for those keywords, this is a totally valid concern, but we can see how Google has changed the practices over the years to navigate this challenge. They started lowering the importance of keywords and raising the importance of user retention or how long somebody stayed on your web site. A site full of unedited AI generated text, but hold a user's attention much less well than a well-written article. And so science with high bounce rates, people arriving in and instantly leaving would have lower search engine rankings. So with Google's May 2021 update, keywords are taking another step backwards and importance in favor of other metrics. All of these steps will help to avoid the problem of keyword saturated AI text lorem people to useless websites for clicks and ad views. All of this means that it's really important to work ethically with AI writing tools to avoid creating content that is hard to understand or useless. On a more personal level, you don't want to be creating content that your readers click away from because it's hard to follow or deviates from your point, from your article. This can be bad for your website's analytics and performance, especially if your blog relates to your work or your business. To combat these issues, you have to have a strategy around your content before you can start plugging prompts into any of the AI tools. In the next video, we'll look at some ways to generate ideas for your blog posts using some different tools. 4. Generating Blog Post Topic Ideas: There are lots of great ways to come up with ideas for post on your blog. It's important to have a plan for what you want, the AI tool you're using to write for you. Otherwise, you're likely to generate something that misses the mark for your blogs goals. There are some great tools out there that can help you come up with post topics that matter to your readers. So let's go over six different ways that you can generate these topics. Number 1 is brainstorming your topics. Now I know this one sounds obvious, but you have to start with what you know about your topic. And this is a great way to expand your perspective of what your blog should be covering. The list that you create off the top of your head can also be a great way to generate more ideas with the prompts that we're going to use for other options. Try to think of keywords that you want to target, maybe questions you want to answer and topics that are relevant to your blog, subject matter and the readers. Number two is to ask your audience, who knows what they want to see from your blog better than people who already read it. If you have any preexisting online presence like social media, you can ask the people that follow you use a pole or a pop-up on your website or send an email to your mailing list to ask. Make it as easy as possible for people to give you their feedback. People often respond better to multiple choice or questions like would you rather see a or rather CB on my blog? This is why having a few untested ideas brainstorm first can be helpful. Number three is do some keyword research. Keyword research will help you to determine what keywords are topics you want to target with your blog. So that's the AI reading tools and they're even ask you for a target keyword to be incorporated into the text that they generate. You can do keyword research using free browser extensions for Chrome, like Keywords Everywhere and Uber suggest which are two of my favorites. There are also subscription sites with many different free tiers that will help you find keywords that you can rank for and see what your competitors are ranking for and generate ideas for topics to hit on based on those specific keywords. Number four is to use a public research tool. More tools are becoming available all the time to figure out what people are searching for online so that blogs and businesses can meet the demand for information. A popular site for understanding what kinds of questions people are taking into search engines for specific keywords is called Answer the Public. This site provides you with a certain number of free searches per day. And they tell you all kinds of variations and related questions that people are searching for on your subject matter. Number five is to browse forums for topics. Another way to explore what questions people are asking about a particular topic is to look through a question aggregator site like Korra or Reddit. This can be a great way to see if there are questions that come up a lot or that people are having a hard time finding an answer for. Additionally, answering these questions on those sites and then linking to your blog post can be a great way to drive traffic to a website afterwards. So keep track of the relevant questions that you find in these forums. And number 6, use AI to generate ideas. Finally, you can actually use the AI tools that we're going to look at to generate blog post ideas. It's important to cross-reference the kind of topics that these tools can generate for you with your own research about what people are actually looking to read online. So consider this a method of creative brainstorming that can generate ideas you may not have considered before come across. The key to making this work is to be very specific about what you are inputting into the AI tool when you are trying to generate some prompts. One way of phrasing your prompt is to the form the AI that the text it is generating is part of a list of blog post titles. For example, if our topic is fitness, then we can type in the following is a list of blog post topics about fitness and see what the tool generates for us. If the tool doesn't seem to want to listen at first, you can kind of help the model find the answer that you're looking for. Start by making a list numbered or otherwise, and put it in an example of a blog post topics similar to the kind of thing you're looking for. This can often be the little nudge that the tool needs to generate some appropriate related topics. Here I'll show you an example of how to use infer kit, which is one of the tools that I mentioned for doing AI writing. So this is basically what it looks like. You just type in this box here. So this is what I mentioned in the example is to type in the following is a list of blog post topics about fitness. And I'll just show you what it does when you just run it like this. So I put number one just to see what it comes up with. So it just kind of giving me all this text that sort of looks like a block was what fitness but doesn't really mean anything. So I just exited that out so we can stay. And I'm just going to try again, but I'm going to put in an example here instead and then just letting it, leaving it blank so that the thing hasn't been run on. So I'll just type in how to stay motivated with a workout routine. And then number two and see what it gives me from that. So from that example, it actually started to realize what I wanted and start to give me all these examples. Now they're not all brilliant. They were kind of, you know, variously, good. So I'm not going to keep all of those, but you can put more prompts into get it more ideas of what you want. Looking for numbered pose like this, like ten ways or five ways to do something is kind of a, another good way to generate blogged type posts. So Sam for the numbering and it started to give me all number based blog post concepts. So that could be a good thing. So you can basically save some of these or you can delete them and just keep hitting generate over and over again. Every time that you give the generator a little bit more text to run off of, it will have a better idea of what it wants to or what you want it to generate. So again, another example. And now after I've put in three of them, I find that the topic starts to get a little bit more diverse in terms of their structure, but they also are more relevant. So at this point I can just keep hitting generate instead of deleting it. And it's gonna keep giving things that may be interesting. Now here at the end you see that it's giving me text that looks like the end of an article saying, did you enjoy this article? And then I'll select these topics which are more headline type things. After you've generated your blog post topic or the question that you're going to be answering. You can then search of breaks that topic down into subheadings, which we'll cover in the next lesson. 5. Create a Post Outline: Now let's go over how you can create an outline for your post. And specifically one that works well with an AI writing tool. Creating an outline for your blog post is helpful for two reasons. Number 1, it helps you to figure out what kind of content your blood needs to contain organized under subheadings, which are really easy for your readers to navigate and search engine friendly. The other isn't to create an outline, is to start generating micro prompts for the AI to fill in. While an AI writing tool technically could just write a post on its own without any guidance. That post probably would take some unintended twists and turns and might not hit on the important points of your subject. So for example, let's try this blob topic in the fitness niche. How to stay motivated with your fitness. And so we can take that topic and turn it into a catchy blog post title, which I've done here, which is three ways to keep you motivated on your fitness journey. I just chose three as an example because it's sort of the bare minimum of examples you'd find in a piece of persuasive writing. Think back to when you learn to write essays in high school, they usually had an introduction, three body paragraphs, and a conclusion, right? So a blog post is structured very similar to that simple essay format. So with our topic chosen, we can use this one here. We can brainstorm three points to make for this article. Of course, you can use the research that you did when generating your blog topic ideas to try and answer it, you can even ask the AI for answers. Sometimes it can provide them depending on the subject matter. However, since you are presumably writing in a niche that you know a thing or two about, it's probably going to be fastest for you to simply come up with some points on your own. So here's what I've come up with three ways to stay motivated with your fitness. My examples are, number one, find a workout buddy and Br2 mix up your routine to keep it interesting. And number three, give yourself healthy and motivating rewards. Now it really quickly, let's think about reader behaviors online. If you're searching for an article online that provides you with information on how to stay motivated with your fitness routine. And you see this list here. What do you think is going to be the most important information that you are extracting from that article. The information provided in the heading is probably most important. The text that follows the idea of final workout buddy, it's probably going to be a little bit obvious and maybe spell out some more information that most people can deduce for themselves. In this case, an AI writing tool can be really handy because you're not going to be worrying about providing very specific technical information in the body text. For the other two points, mixing up your routine and finding healthy rewards, you're going to want to find some examples to contextualize your statements. In this case, the AI may or may not provide you with examples when it writes for you. This is where it becomes important to edit your generated text and make sure it includes relevant and applicable information. This is an example of where an AI writer may not be able to be useful. So on that topic, what can AI, right? So what AI or any writers are very good at right now is abstract or conversational texts that broadly describes a concept or discusses a topic with a moderate level of expertise. They're also very good at writing fiction. I will say, you know, very good is relative, I suppose, but they're not bad at writing fiction. What they are not so great at would be a step-by-step technical blog post. So for example, how to set up a printer. That information needs to be very specific to the device that you're working with. And an AI would not currently be able to provide you with accurate enough information. That may be something that is possible in the future. But right now there is not a lot of ability to generate very specific technical steps. The same will be true for running a recipe. The recipe that an AI could write could perhaps be delicious, but odds are pretty good. The proportions IP off because it doesn't have that sort of conditional logic based on recipes. So there are a few other pieces of information that you can include in your airline. If you are linking to a product and other source or promoting something, you can decide where in your article that information will be included. This will be important for the editing and rewriting stage of your article. There may also be other parts of your blog that can be preplanned, like maybe a signature that says, I'm saying that you always use or a link bank at the bottom of your post. You can also do some pre-writing here, including anything that you know for sure you want to stay in the article, the AI texts that you generate next can fit in around this text and you can even add in your own words to the generator as you go to coach it to write more and your natural language style. So in the next lesson, we're going to start plugging some of these prompts that I made up into the generator to see what we get. 6. Feeding Your AI Prompts: Now let's start feeding some prompts into the AI writing tools to generate some text. I'm going to be continuing to use infer kit for this project. But there are other tools as I mentioned, and there are also other ones that I like to use. It just depends on the type of project I'm doing. But for this case, this one is very straight forward. So again, this is the infer Kit interface and you just put the text in the text box there on the sidebar, you can decide how many characters is going to generate per generate click. You can also add in things like the category and keywords if you want to make it more specific. So to get started, you want to provide the program with some contexts for the kind of thing you want to write. So I'm going to include the title of our blog post, which is three ways to keep you motivated on your fitness journey. And after we get the title and I'm just going to put the blog post topics that we already came up with, which we're finding workout buddy, mix up your routine to keep it interesting and to give yourself healthy and motivating rewards. So once these are in, I'm gonna go in and give it a little bit more text to work with. So just trying with the introduction, I'm going to put in some, some texts that just kinda sounds like it would be good for the introduction. So here I'll just put it in a sentence. Everybody wants to stay on track with their healthy routines, but it can be difficult to find the motivation to keep going to the gym and making healthy lifestyle choices. So this will give the idea of the kind of flavor I'm looking for in this text. And here's what it spat out. I didn't really like it, so I'm not going to keep it. But that's the beauty of AI. You could just keep generating to something actually comes out. In this case, this text here is very specific to a time of year, but maybe that's the context we went for the post. So you can just trim it up, keep it you want, delete what you don't, and then move on to the next section when you are ready. Now we'll just go down to the first topic for this bug host. Again, I'm going to give it a little starter Sentence just to kinda figure out what it is I want it to generate or to follow up creating texts for. There was my first sentence and you know, just kind of putting in some contexts. Of course, this is where your previous research is going to come in handy and any knowledge you have on the subject matter. So once I generate it started giving me this information that sounded really legit. It's about this doctor who is not a real person. And then there's a quote and it just gave me some stuff that was pretty workable. So I just kept pressing generate to see what else it would give me. And then you can trim it out and take what you want. So it ran to a point where it actually decided it was done and gave me a second which I didn't want, so I deleted it and then moved on to the next section. So again, giving it a little sentence that just generates what exactly we want this point to be about. I also find putting questions in there like how do you keep it interesting? And then it says, make it a game. So if you put a question, often the AI will answer you. And that's a good way to kind of direct the way you want it to start writing. So in this case it is giving me lots of examples. And then again, it started giving me the third when it felt like it was done and it just kinda kept doing that. So I figured, okay, this is a natural stopping point. I can add more text in later. Moving down to the last point, I will generate some sort of legs statement again, giving yourself something to look forward to and the other side, if your workout can be a great way to push forward, push through those difficult days. With this kind of starter in place, the air will give me some more relevant texts. You can kinda see the pattern here where it's going to fill in texts that responds really well to the prompts that I put in. Which is why having the very concise outline like this a lot easier. Otherwise, the AI can start taking this in directions that are completely different. And here didn't look at me what I wanted. So I changed to a question format again, and I started to do a little bit of a better job and actually start to generate some suggestions for what the points could be under this heading them in a lot of sense. It also started to go on to make a fourth which I didn't really want. But I mean, it could, if you wanted, if you saw it, maybe one that was really helpful or relevant to your topic. Again, it could inspire you to add some additional text or make another point for your blog post that you like. At this point I'm just going to go to the end of the article and start to add some sort of conclusion statements that you can. You can also do this just in your Word document, but I like to kind of do it right at the end and just see if this generator will give me a few interesting thoughts to wrap up my article. And sometimes at the end it'll, again, it'll start to format it like the end of a blog post. Because remember, a lot of the Internet is the source material for this generator. So it knows the structure of a blog often ends with see more. Or if you liked this article, you might like these. And then you can wrap it up and just copy and paste this text into your word processor. In the next lesson, we'll start to clean up this text and turn it into something that's worth publishing. 7. Editing and Rewriting: Once you've generated some text for your article and may note of all the additions you want to include, links, referrals or products. You can start to clean up that text. I usually like to start with making sure all of my text is formatted in the same font and size, often copying and pasting it can mess up the formatting. What are the tricks that I like to use is to change the color of all my computer-generated texts. And maybe I usually put a blue so it's different so that I can easily tell what has not been written by me. This makes it a lot easier to know what sentences you need to be really thorough and proofreading to make sure they sound like something that makes sense that a human being would say and bought something that's generated or maybe a little bit confusing or off-topic. If there are any sentences that don't make logical sense, just try rewriting them or changing some of the words to sound more conversational or whatever the tone of your blog is, one of the best ways to edit computer-generated texts to sound more natural. And this is also something I used to teach my students about editing in general is to try reading it out loud. This makes it really easy to tell if it sounds normal like natural human speech or writing as opposed to something a little more mechanical. If you find that this section is missing some information or you want to make a particular section longer. Having some texts, they're already makes it a lot easier to add in your own words. This isn't really good way of powering through writer's block in general, if you don't know what to say, ask the computer to say something and maybe something that comes out next is really good and you can roll it that. Of course, you can always keep adding in more and texts. But at this point again, you might find that it's just faster and simpler to add your own words. This is wrangling your texts back into the AI tool. However, if you do decide to go back and try to generate more, you can paste in your whole article or just the section that you're working on and generate from there. The more logical and edited text that you can provide as a source that it's going to cross-reference with the corpus that it has on the backend. The better job the program could do and creating more text. When you're editing your text, you can start to add in some of the keywords that you are trying to target for this article. So for example, if you're writing your own fitness and you see the word workout come a lot. But you are trying to write specifically about at home workouts that you can go in and just change the keyword there from workout too at home workout to make sure that you are showing up for that keyword. Even though I said it is not as important as it used to be with search engine ranking. Keywords are still very important in terms of understanding what the content of your article is about and how to index it. Try to find organic places to weave in any promotions that you're adding to this post. So let's say for example, you are promoting a certain brand of frozen smoothie is in the Post. It would make sense that this would be included in the positive rewards section that I created. You could position these new these as a great tasty treat after you've had your workout, you would want to reread the section and figuring out where the best place would be for that product placement. You can also add in your photos and images to go with your blog posts. This would be done when you are just uploading the blog to your WordPress site or wherever your block is hosted. And before you wrap up, you'll just want to add in any of the extra pre-written text, like a closing statement or social links, then you should have your blog post texts ready to go and share with your audience. So let's take a look at some texts that we generated for our example blog post and how you could edit them. So here's a few things that you could improve on. So the sentence about the AI gave me was, chances are you have a similar training schedule and a and fitness level. You'll know what to expect in the exercise session. So that's okay. You clearly you can tell this belongs to the first one about finding a workout buddy. But the sentence doesn't really fit into the context of the subject heading. Although it's on theme, it doesn't grammatically work and flow with everything else that was included. So this is how I improved it. The edited statement. You can see I wrote when selecting a workout buddy, look for someone with a similar training schedule and similar fitness level. This will help you to know what to expect from your exercise sessions. So what I did here was I made the subject clearer by specifying something or workout buddy. And I added some contextual language around the second sentence to make it easier to skim and understand. Here's another example. So this paragraph that the AI generated reference a fictional doctor. So it reads, having someone to keep you accountable and your work has this really helpful. But how do you find the right person to become your workout buddy? That's classifying texts. Here's what Dr. Gregor Moscowitz, a sports psychologist who specializes in goal achievement, has to say about finding the right workout buddy. And it went on to provide a quote. So I don't think this is a real person generally these peoples and their quote sound very real when you use AI, but it's just blending people's names who maybe had a similar role in other texts. So we don't actually want to be citing fictional people for our blog post. That's not really cool. So this needs to be modified. So here are a few suggestions as to how we could modify this paragraph for our blog post. You could find an actual expert and find out what their opinion is and swap their advice in. You could find an academic paper on this subject if you don't want to talk to another human and quote them, you can find academic papers on Google Scholar, for example, is a good place to look. You can find another online source to quote another website and your blog that you could link to quote there. Or you could simply just change the language in this section to avoid using a direct quote and speak more broadly on the point. So now that you've seen how to use this type of tool to make a blog post easier and faster. Let's take a little look at a few other ways that you might want to use it. 8. Other Uses: Ai writing tools have a lot of potential to make so many types of writing tasks a little bit easier before we wrap up this course, here are a few other ways that you might want to use some of the methods that we've seen today. Number 1 could be in writing email newsletters, it can be hard to write a really strong marketing copy for newsletters, especially because they are being sent out on a regular basis. And it can be tricky to reuse the content that you write in them. You can use an AI writing tool to help compose those emails, combined with some persuasive marketing reading strategies. There are also specific AI tools out there that generate e-mail newsletters or just emails in general to help with time management. An example of an email generator, Time AI. Number two, you could use an AI writing tool to help with writing social media captions. If you struggle with what to say on Instagram or Facebook, you could try and air writing tool to inspire your captions. You can use a general tool like I've been using here, or try wonderful many AI copywriting tools out there that allow you to input the name of a product or a tagline that you want included in the text. You can also use AI to write product descriptions if you run an online store using a very similar method. But of course you want to make sure that's very specific to the product. So lot of proofreading on that. One. Number three would be writing press releases, please. This can sometimes sound like that being written by computers, even when they are written by humans. If you need some informational texts written up that has specific keywords or info, but a product or event. You can either use a generator specifically for this task or just use one of the generic ones and edit the text to include the key points for your press release. Now there are so many ways that AI writing tools are being used online already. Like guarantee you that you read AI generated texts on an almost daily basis without realizing it. By harnessing the power of some of these emerging tools. Not only will you be able to write effective texts faster, but you'll be able to take advantage of some of the even more advanced tools that are coming down the pipeline in the next couple of years with less of a learning curve. Let's hope that's motivating. I hope that you found this course useful. And if you want to leave me a review for the course, it would really mean a lot. I've read them all and it makes my day. The course assignment is outlined in the assignment section of the course player, but you'll find a template there that will help you to outline your blog post. You prepare it for working with an AI tool if you're choosing. This should help you get started brainstorming your topic and preparing the necessary steps to generate some content as I've shown you in this course. And if you are interested in other topics around writing, publishing, art, or technology, you can check us out my other course offerings. Good luck with your projects and thank you for joining me on this AI adventure.