Over the past few weeks, it seems like AI has become the center of attention. My LinkedIn feed is flooded with thought leaders, tech evangelists, and marketers sharing insights on how artificial intelligence will transform the way we work. Similarly, on Instagram, AI platforms promising to “unlock creativity” and “enhance daily tasks” have started targeting me (looking at you, Jasper).
This heightened focus on AI is justified, as recent developments have yielded impressive results. Furthermore, the number of new AI tools and services hitting the market is growing exponentially, with estimates of 100+ weekly releases. What stands out is the increasing accessibility of these tools. Many AI developers are offering free or user-friendly formats that appeal to less-technical individuals. This trend democratizes AI’s capabilities, opening up opportunities for people across various industries to experiment and innovate using AI. A prime example is OpenAI’s ChatGPT.
For those unfamiliar with ChatGPT, it is an AI language model created by OpenAI, designed to converse with users in a natural and human-like way. It uses deep learning techniques based on the transformer architecture, specifically GPT (Generative Pre-trained Transformer) architecture, allowing it to understand and generate natural language text in response to user inputs.
Since its release, ChatGPT has attracted college students, developers, marketers, and many others, thanks to its easy-to-use interface and chatbox. Users with little to no knowledge of machine learning can generate interesting results, experiencing a conversation similar to messaging a colleague, except with access to vast amounts of internet data.
The key lesson is not to underestimate the power of good user interface design. By humanizing and simplifying the experience, OpenAI has driven huge volumes of sign-ups, allowing users to transfer existing behaviors for instantly usable results.
However, the platform has limitations. ChatGPT’s training data only goes up to 2021, making it unable to offer responses about current events. It sometimes “hallucinates,” presenting completely fabricated responses, and its guardrails to prevent dangerous responses aren’t perfect, often resulting in unanswered questions or overly cautious, repetitive stock responses. Or, as one New York Times writer found out when trying the new GPT powered Bing app; amusing and, somewhat, unsettling confessions of love and a desire to be destructive.
Adding to the AI buzz, OpenAI recently released GPT-4. Although the ChatGPT interface remains largely unchanged, GPT-4 is multimodal, meaning it can see, listen, and has advanced reasoning capabilities. Early results are astounding, as it can ace exams, solve problems with incredible accuracy, and produce stunning results for tasks like doing taxes, suing cold callers, creating video games, and building websites.
Currently, only ChatGPT Plus subscribers can access GPT-4 for text chat, and the new image analysis functionality is withheld due to potential misuse concerns. However, the waitlist for the GPT-4 API offers significant opportunities for developers, further empowering creativity and new possibilities with generative AI in the near and long term future.
In the meantime, some marketers are leveraging ChatGPT to enhance and augment their abilities. Here are three great examples:
#1. Julian Cole – ChatGPT Guide for Strategists
In this guide, Julian outlines four ways strategists can incorporate ChatGPT into their process. I love how through the use of imaginative prompts, this guide shows how to turn ChatGPT into a tool for helping generate new insights, explore existing customer perceptions and even find the best words for your strategic presentation.
#2. Chase Diamond – 25 ChatGPT Prompts to 25X Your Email Copywriting Results
Here, Chase shares a series of text prompts that can be updated with details about your brand, product benefits and your target consumer. When inputted into ChatGPT they will generate different email subject lines and body copy based on different messaging frameworks.
#3. Ksenia Newton – A Social Media Manager’s Guide to ChatGPT
In her article for Brandwatch, Ksenia provides advice on what to do (e.g. define the tone of voice & give context) and what not to do (rely on the word-for-word answers) with the tool. In addition, she shares a number of example prompts that will help SMMs generate different social media posts.
The common thread in these examples is the emphasis on precise input quality, or in other words, a well-crafted brief. It’s a familiar equation; quality in equals quality out, so it makes sense that we invest time in the upfront to test, refine and perfect the prompts to get to more interesting and quality responses. Additionally, each encourages users to continuously review and critique the output, ensuring it meets the brief’s requirements. These actions must be performed by humans, not machines, at least for now. This dynamic indicates how GPT and other AI tools will shape our future as marketers; we’ll become more efficient, effective, and creative in briefing these tools to empower better insights, strategies, and activations, rather than relying on them for shortcuts to output, which could lead to bland, repetitive execution devoid of originality. As Julian Cole puts it, “GPT requires you to play creative director. Not artist.”
So, back to my newsfeed, where the discussion continues at pace; will AI put us all out of our jobs? The consensus so far is ‘no.’ And from my point of view, I agree, at least for now. While AI tools like ChatGPT, Dall-E, Jasper, and others will undoubtedly change the way we work, they are meant to augment our abilities, not replace them. It’s crucial therefore that we invest in and stay up-to-date with these new tools because, while an army of Skynet machines in Converse and plaid shirts may not replace us, a human using AI might!