My Experience Writing With AI in 2023

Artificial intelligence seems to be everywhere these days. Beyond its buzzword status, the potential of AI to complete various tasks has generated both excitement and skepticism—and writing with AI is no exception.

As someone who spends countless hours wrestling with words and ideas, I’ve found AI to be a helpful tool that’s transformed the writing process. However, like any tool, it comes with its own set of frustrations and concerns. Whether you’re all for it or still skeptical, one thing is clear: the significance of AI is only growing stronger. So, let’s dive into the current pros and cons of teaming up with AI for writing.

The pros of writing with AI

A fresh perspective

AI is great for when you need a fresh perspective, which is especially useful when it comes to tedious or repetitive writing tasks. Consider the process of crafting hundreds of product descriptions or social media captions for a brand. It can be quite a mental workout to come up with a dozen different iterations of the same information.

In these cases, AI can step in to provide diverse alternatives that not only save time but also inject a fresh energy into the creative process. Its ability to generate multiple variations effortlessly allows us to consider different tones and styles and push the boundaries of creative expression.

AI can inject a fresh energy into the creative process for writers.

This aligns with a broader discussion in the Harvard Business Review on how generative AI can “promote divergent thinking” and “challenge expertise bias.” Ultimately, by being exposed to the diverse content generated by AI, creators can be inspired to “think beyond their preconceptions of what is possible or desirable.”

Overcoming writer’s block

AI has also served as an effective antidote to writer’s block—an all-too-familiar hurdle in the writing process. Countless AI writing tools market themselves as the ultimate solution or cure for getting rid of this creative obstacle and I, for one, have found it to be true.

When I’m struggling to find the right words and express a particular idea, it’s been helpful to throw that raw, unrefined concept at AI and have it quickly suggest various phrases or options. This helps me break through the mental roadblocks and get my writing flowing again. It’s a straightforward and efficient tool, similar to consulting a trusted friend when words temporarily escape you.

A personal research assistant

Additionally, AI can be a huge help in the research process, making it much more efficient. By feeding it an initial idea, AI can provide concise summaries of relevant resources, which can help writers quickly explore different angles and hone in on specific themes. This speeds up the initial stage of content creation, offering a refreshing alternative to spending hours sifting through countless resources for the desired information.

It’s important, however, to think of AI as a starting point for research, an assistive guide to help get us in the right direction. AI is powerful but it’s not without its drawbacks, and those writing with AI have to be careful not to put too much faith in what it generates. After all, ChatGPT has an ever-present disclaimer at the bottom of its page: “ChatGPT can make mistakes. Consider checking important information.”

The cons of writing with AI

Unreliable accuracy

One of the most glaring issues with writing by AI revolves around accuracy—or the lack of it. The frustrating reality is that AI can be quite unreliable, with the tool known to pull “statistics” and “facts” seemingly out of thin air.

Even with its disclaimers, the way that AI confidently presents information can be misleading. If not rigorously fact-checked by responsible users, there’s a genuine concern over the potential of widespread misinformation.

To prevent the rise of AI-induced misinformation, we must strengthen fact-checking institutions.

While researchers argue that these concerns are largely overblown, they do acknowledge that the current rarity of misinformation (at least, in wealthy, democratic countries) is “thanks to the hard work of professionals—journalists, fact checkers, experts, etc.—and to the norms and know-how that have developed over time in these professions.” Consequently, to prevent the rise of AI-induced misinformation, we must strengthen those institutions.

Lack of transparency

Another concern is over the origins of the information AI generates. While search engine AIs may occasionally offer source links, general-purpose AIs are typically unable to provide proper citations. This lack of transparency casts doubt on factual claims, makes it hard for users to find verifiable information, and raises ethical questions about using proprietary ideas without attribution.

As McKinsey reports, “inaccuracy, cybersecurity, and intellectual property infringement are the most-cited risks of generative AI adoption.” These concerns underscore the need for “a human in the loop” who can verify and interpret AI’s outputs at this stage of its development.

Language and creativity concerns

Language is another area where AI reveals its weaknesses. Unless explicitly directed to avoid them, the tool tends to overuse certain phrases and wording. This raises concerns that content will all start to look and sound the same, with a risk of losing the unique voice and creativity that human writers bring to the table.

This limitation can be addressed by writers who understand the nuances of language and are equipped to improve the generated content. However, for users with less writing experience, the risk of most content becoming stale is a valid concern.

Acknowledging AI’s benefits while also making sure it doesn’t stifle human creativity and diversity is a balancing act.

In particular, there’s a genuine worry, particularly for language lovers like myself, that an overreliance on AI could lead to creative stagnation. On one hand, if society leans too heavily on AI for creation, we may end up with bland, formulaic content that lacks the authentic voice most of us value.

On the other hand, an aversion to AI-generated content can stifle human creativity. Preconceptions about AI’s style may confine writers, preventing them from using normally acceptable figures of speech, as certain expressions become unfairly associated with AI. For instance, within the past year, I’ve heard comments about how similes and metaphors come across as too AI-like. Ultimately, it’s a balancing act between acknowledging AI’s benefits while also making sure it doesn’t stifle human creativity and diversity.

Striking a balance

Unlike automation technology of the past, which primarily impacted physical labor, AI is now having the most significant impact on knowledge work. Researchers estimate that AI’s performance will be on par with the top 25% of human workers by 2040. In some cases, that’s 40 years earlier than previously predicted, sparking major curiosity about the future of AI.

AI can provide efficiency and inspiration in content creation, but it also demands careful consideration and oversight.

McKinsey argues that AI could significantly boost global GDP by increasing labor productivity—so long as workers are given the resources to learn new skills and shift to other work activities or occupations. At this moment in time, though, while AI can be a huge help with certain functions, it’s crucial to be aware of its limitations and tread carefully.

For writers,  AI can provide efficiency, inspiration, and a fresh approach to content creation, but it also demands careful consideration and oversight. The responsibility lies with all users to fact-check, inject creativity, and give credit where it’s due. Let it complement our skills, not replace our critical thinking and human intuition. All in all, the optimal approach to writing with AI—for now—is to embrace it as a partner, not a substitute.

___

Blog illustration of primitive script keyboard commissioned by Anne Mieth using generative AI