From AI Authorship to AI on the Shoulder: 5 Tips for Using Generative AI

In many ways, 2023 was the year of GenAI.

ChatGPT, now a household name that is practically synonymous with generative artificial intelligence, made its debut a month before the beginning of the year. And throughout 2023, companies like Google, Anthropic and OpenAI continued to announce improvements in the processing power and capabilities of their products. Suddenly, organizations across industries were expected to have detailed and forward-looking plans on how to use generative AI to transform their businesses.

And yet, for business leaders and creative professionals who have attempted to use GenAI in their work, the experience has often been frustrating. This is partly because GenAI has at times been discussed as a magic button. People think that you can ask ChatGPT to create a marketing plan or write a white paper, and the tool will spit out the business deliverable, fully formed.

But we’re starting to see a recognition that GenAI, at least in its current form, is more useful when humans treat the technology like a collaborator. Organizations such as the software company Autodesk have begun to refer to this model as “AI on the shoulder” — suggesting that GenAI solutions often act less like a sole creator, and more like a highly efficient assistant.

Here are five examples of how businesses might use this model to create value with GenAI today:

1. Business Analysis
Warren Buffett is reputed to spend up to six hours a day reading, ingesting thousands of pages of information about Berkshire Hathaway’s current and prospective investments. Far be it from me to suggest that the Oracle of Omaha tweak his process after decades of success. But for us mere mortals, AI can jumpstart the process of business analysis — summarizing hundreds of pages and finding patterns in large data sets in a matter of seconds. Now, if it were my job to dole out investment advice or allocate capital, I would actually read the report myself as well. But by letting AI take a first pass, business analysts can set themselves up for a deeper dive that focuses on the most important points of a report.

2. Proposal Writing
As an academic, I confess that I’ve started using GenAI tools to help me write my grant proposals. Does this mean that ChatGPT is doing my work for me? No. But I can use a large language model to brainstorm ideas, assist with budget and resource planning and write first drafts of proposal sections. It still takes me a full day of work to write a 10-page proposal, but it used to take me two weeks. With the time I save, I can conduct research, provide students with more targeted attention or simply write more proposals. Without my input, ChatGPT would only be able to create a generic proposal that would never help me secure funding for my research. But without ChatGPT, my proposals would take me 10 times as long to write.

3. Translation
My father, who works in the humanities, constantly razzes me about the quality of my written Spanish. Since all of my higher education has been in English, I sometimes struggle with the finer points of Spanish grammar and punctuation — particularly with where to place accents on vowels. I wouldn’t necessarily trust a large language model to translate an entire paper for me. However, a tool like ChatGPT does a great job of checking my own translations for me and making sure they meet my father’s high standards! I still need to closely direct the GenAI solution, instructing it not to change my tone, voice or style, and to focus solely on fixing my linguistic typos. But it can do that pretty well.

4. Coding
I like to code live in the classroom. Even though I’m the instructor, it can still be unnerving to write code in front of a classroom full of brilliant MIT students. Until recently, I spent an entire day preparing for these lectures, along with a couple more hours on the morning of class. Today, I use tools like Microsoft Copilot that provide me with a safety net in case I get stuck or make a mistake. I still prepare carefully for my courses, but my morning-of work has shrunk from more than two hours down to around 30 minutes.

5. Content Creation
If you ask ChatGPT to write an article from scratch, it’s probably going to come up with something fairly generic. But if you instead use the tool to help spark ideas and improve your own writing — at times leaning on it almost as an all-purpose autocomplete feature — you might find that you’re able to produce a blog post or article in half the usual time. Over the past year, we’ve gone from “write me a blog post” to “help me make this blog post better.” This means that people are increasingly looking for GenAI to help them with their creative work, rather than to replace human creativity. That’s a big mindset shift — and, I would argue, a big improvement, as well.

Source: Reworked