ChatGPT’s meteoric rise to popularity has dominated the news in the last six months. Within one week of the tool’s launch, many media outlets were praising the software, with The New York Times admiring its “brilliance and weirdness” and TechMonitor saying it’s a “game-changer for businesses.” But soon, it became clear that ChatGPT, though revolutionary, is not without its issues, raising debate about whether you can trust ChatGPT to create content for business. This blog delves into the debate, looking at the good, the bad, and the ugly of using ChatGPT to create content.
According to its own description, ChatGPT is “a large language model trained by OpenAI, based on the GPT-3.5 architecture. It is designed to be able to engage in conversations with humans and generate responses that are fluent and coherent, using natural language processing techniques. ChatGPT has been trained on a massive amount of text data, and is capable of understanding and generating responses in multiple languages. Its abilities include answering questions, providing recommendations, and engaging in general chit-chat.”
Reddit is a gold mine of examples of ChatGPT getting information wrong. The article below shows a user asking how many country names begin with V. ChatGPT’s response is that no country names start with this letter. When we asked ChatGPT the same question (see screenshot below), we initially got a different response, saying that there is only one country name that starts with V: Vatican City. This is still wrong, as we’re sure many Venezuelans, Vietnamese, and Vanuatuans will attest.
We asked again and received the answer that no country name starts with V. Admittedly, ChatGPT might have used the full country names (the Bolivarian Republic of Venezuela, the Socialist Republic of Vietnam, and the Republic of Vanuatu, respectively) in its first answer because, when we queried these countries, ChatGPT replied that it had made a mistake. Still, all three countries are better known by the name that starts with V, so it’s curious that ChatGPT doesn’t mention them. Clearly, it’s always best to fact-check AI-generated content before you hit Publish.
In early April 2023, ChatGPT’s reputation was significantly tarnished when USA Today released an article stating that the AI software, citing a 2018 Washington Post article for corroboration, had falsely accused Professor Jonathan Turley of having sexually harassed female students on a class trip to Alaska. The problem was that no one ever accused Professor Turley of sexual harassment, he never attended a class trip to Alaska, and this Washington Post article never actually existed – ChatGPT had invented it.
This example is, obviously, very extreme, but it does raise the question about the legal issues surrounding using ChatGPT to create content. Forgetting countries is one thing, but accusing people of crimes they haven’t committed could leave a content creator open to lawsuits.
Here at Proofed, we’ve talked at length about the importance of creating grammatically accurate copy (you can see one of our blogs on that topic here), but it is especially important to check content that AI has created. Developers use information from the internet to train the software, so the software’s accuracy depends largely on the datasets the developers use. Languages are so complicated that it’s impossible for a machine to know every grammatical rule in every situation. Chat GPT itself says that “there may be occasional errors or inconsistencies in its responses, particularly when dealing with complex or ambiguous sentence structures or idiomatic expressions” (see screenshot below). The easiest way to overcome these issues is to employ an editor and proofreader to check your writing for grammatical and spelling mistakes. Here are our six tips on editing AI-generated content.
AI-generated art swept across social media in 2022, leaving many artists horrified that AI art generators were copying their art. In an article entitled “The Dark Side of AI Art,” Simon Tolcheva notes four main issues with using AI art generators: the harsh terms in the fine print, the potential issues with copyright, the question of whether AI-generated art is stealing from artists, and the promotion of harmful stereotypes or biases.
AI writing tools have also faced criticism about the software violating copyright and stealing from writers. The second aspect of this criticism is somewhat unfair. AI does not directly use writers’ work and claim it as its own, but it does use anything published online in its training. And there are issues with copyright. Because ChatGPT and other AI writing tools are so new, it’s not clear who owns the copyright on a piece of work: the person who input the prompt or OpenAI. This issue will likely be resolved over time, but for now, it’s something to consider when you’re using ChatGPT.
You might be tempted to avoid ChatGPT like the proverbial plague after everything we’ve just said, but we would be remiss not to discuss some of the benefits of the software.
In 2022, Google’s Search advocate, John Mueller, announced that Google’s algorithm would penalize AI-generated content, much to the dismay of many online creators who used AI to help them with their SEO. So, rather than helping, ChatGPT hindered the SEO efforts of these creators. However, a year later, Google revised its stance, saying its issue was with creators using AI solely to rank higher in the search engine and not to create quality content. It seems that the unfair notion that AI-generated content is bad for SEO has stuck.
If you’re anything like us, we’re sure you don’t have enough hours in the day to do everything you want or need to do. This is where AI really becomes a vital tool. AI can help you streamline your workflows and do the more boring and time-consuming tasks. For instance, you can use ChatGPT to create outlines from which your writers can work, something that always takes longer than anticipated, but this has the added benefit of creating consistency. ChatGPT can also write your social media posts, which your social media team needs to tweak only, freeing them up to do the bigger tasks. These are just two examples of how you can use ChatGPT to streamline your workflow.
AI has long been heralded as something that can assist us in being more productive, and that’s exactly how you should use AI. Replacing humans completely with AI leaves you open to accuracy, legal, and ethical issues, but using AI in conjunction with your human teams can help you create more content more consistently – a massive win for any business.
Knowing how to use AI to assist your content production processes is challenging, so if you’d like more information about using AI in content generation, you can read our blog here.
Reach out to us today if you need help editing your AI-generated content and getting it publication ready.
Using AI to Generate Content?
Let’s talk about the support you need.