Article's Content
Article's Content
In his book “The Coming Wave,” Inflection CEO Mustafa Suleyman talks about how truly disruptive innovations—powerful general-purpose technologies—produce a wave-like effect in society. From the mastery of fire to the emergence of agriculture to the practice of writing, there have been only a handful of such technologies in the whole of human history.
If recent AI trends are any indication, Suleyman argues we’re in the midst of another seismic wave.
As we step into 2024, the tech landscape is being reshaped by the relentless advancement of generative AI technologies. These innovations are not just tools for efficiency; they are becoming central to strategic planning and creative execution. Marketers, creatives, and software developers are now at a pivotal juncture where embracing AI can significantly boost their campaigns, outreach, and customer engagement.
This blog post explores the emerging generative AI trends marketers must monitor as we approach another new year in the era of artificial intelligence, including:
Let’s get started.
Generative AI Trends to Monitor in 2024
These are undoubtedly exciting times—particularly for marketers, creatives, and those who work with a computer during their day-to-day lives. Generative AI has opened the door to myriad benefits and plenty more that are yet to be discovered.
If you ask the enthusiastic analysts at McKinsey, the productivity increases AI brings could add as much as $4.4 trillion to the global economy annually. This projection includes generative AI’s impact in over 60 enterprise use cases across departments like sales, software engineering, marketing, customer success, and more.
The global marketing industry could see an additional $400-500 billion in added value across cost savings, new revenue opportunities, and optimizations.
With that miraculous number hanging in the air, let’s look at some of the major trends to keep your eyes on as we approach another year in the AI era.
The Rise of Content Creation Centaurs
Applying artificial intelligence in the workplace is a relatively new concept for marketers and creatives, but it’s something chess enthusiasts have been doing for decades.
IBM’s Deep Blue beat chess world champion Garry Kasparov back in 1997. But rather than discouraging players and destroying the industry, this monumental event propelled a new generation of AI-human chess teams known as centaurs.
These cyborg-esque teams actually perform better than an individual human or bot player. The player brings creativity and intuition to the table, while the computer brings unmatched analytic capacity. It’s a winning combo.
Well, a new breed of centaurs is emerging in the content creation landscape.
Anyone well-versed in the use of generative AI tools now has a sizable advantage over anyone who isn’t. Whether you’re a marketer or sales rep, crossing the chasm to adopt these new tools will help you pull ahead (provided you use them correctly). Much like you grow accustomed to the communication styles and tasks best suited for your colleagues, content centaurs learn the ins and outs of large language model-ese.
Interestingly, recent studies find that soft skills are important for improving the outputs from generative AI tools. Researchers found that using “EmotionPrompts”—prompts that appeal to the emotional intelligence of generative AIs—can yield relative improvements of 8-11%, depending on the task.
The little AI interaction below isn’t exactly what the researchers had in mind, but I’ll submit it as my contribution anyway.
While this piece focuses on content creation, the phenomenon certainly isn’t exclusive to writing and chess. As generative AI penetrates more and more of our everyday devices, expect to see a lot more centaurs out there.
Tech Workers Reach a New Level of Productivity
AI technologies are transforming the role of information workers across the enterprise landscape. According to McKinsey, sales and marketing personnel lead the way, with 14% using generative AI regularly at work. Not far behind are product and service development teams at 13%.
The fact that marketers and sales teams are leading the way in enterprise AI adoption isn’t a surprising trend, considering these departments create a significant amount of text-based content. The same McKinsey report found that the most common uses of generative AI in the workplace include:
- Creating the first draft of text documents (more on this in a second)
- Summarizing and searching text documents
- Creating personalized marketing assets
- Identifying trends in consumer behavior
- Drafting technical documents
- Creating new product mock-ups
This trend projects to grow substantially as more and more people realize the productivity benefits of AI. This next year should also see the workplace usage numbers increase across other key lines of business, like operations, risk, strategy, human resources, and finance.
The National Bureau of Economic Research puts average productivity increases from AI between 14-34%, depending on the task. And for all we know, that’s an underestimate.
Tech and media companies are particularly well-positioned to see above-average performance returns on AI integration. The products and services in this industry are almost exclusively digital and store vast amounts of data that large language models find quite tasty. It’s why we’re seeing AI expenses trend upward in company budgets as brands look to apply these models to reduce costs, bolster products, and unlock growth opportunities.
Don’t expect most tech companies to be among the ones calling for a pause on AI development.
Prompt Engineering Is a Must-Have Skill
The prompts you feed to ChatGPT, Jasper, or other generative AI tools should, in the majority of cases, contain three components:
- Context: A description of what you need the AI to help you accomplish
- Input Data: The copy, data, or images you feed the AI to inform it’s response
- Output Indicators: Formatting specifications for what the output should be
The practice of tweaking and adjusting these components to extract better results from AI models—known as prompt engineering—is essential for realizing the full potential of artificial intelligence. It’s a skill long known to programmers that’s quickly becoming a must-have skill in content creation with LLMs.
Effective prompt engineering can significantly enhance the quality of AI-generated content, from producing more relevant copy to matching a specific voice or style. Experimenting with different prompting styles, particularly multi-stage, chain-of-thought prompts, is a great way to build familiarity with these tools.
Prompt engineering is a nascent field in the context of large language models, so keeping tabs on the latest developments is key. Researchers at Vanderbilt University’s Department of Computer Science recently curated a catalog of prompt patterns.
Prompt Patterns are codified approaches to formatting and customizing prompts so the interaction with an LLM delivers a certain output category.
For instance, using ChatGPT to create a recipe, template, or persona requires an output customization prompt. The Vanderbilt researchers identified 15 different prompt patterns that you can apply to produce better prompts.
As these models get more powerful, it’s easy to see how optimizing prompts can give marketers, content creators, and other demand-side roles a sizable advantage. Even the most bare-bones prompts can help expedite the brainstorming stage for everything from blog posts to customer personas:
Diversifying Your AI Stack Becomes Critical
The recent events surrounding OpenAI have put the company’s future—and, more importantly, that of ChatGPT—into uncertainty. It’s also a reminder that you can never put all your eggs in one basket, especially in tech.
If you (like me) have spent the lion’s share of the last year experimenting exclusively with ChatGPT, it’s time to change that.
This isn’t me ringing the alarm bell, saying it’s time to ditch OpenAI and go all in on one of its competitors. It just means that you should, when possible, test out other tools. The lead position among generative AI tools currently belongs to ChatGPT, but that can quickly change if the price hikes, a competitor has a major breakthrough, or the talent behind it moves to another company.
Being a jack-of-all-trades is much more advantageous than being a master of one in an industry where one can become none overnight.
Many of the current pieces of your tech stack—Notion, Salesforce, Canva, and more—already have AI value-adds through APIs and integrations. Even taking advantage of free trials to test-run the models from other developers is instructive. Every new context where you apply AI will help you build the skills to be multilingual in the AI realm.
We’re All (Kinda, Sorta) Artists Now
Creating the title image for this piece—what I’m now calling the CMO Cyborg—took me all of one minute and a $20 monthly subscription. What would have otherwise taken me an absurd amount of time using expensive graphic design tools and ultimately resulted in an unusable mess is now incredibly easy.
This easy:
While many are (rightly) focusing on the text generation, manipulation, and analyzation capabilities of generative AI, the ability to create content is also growing to include more dynamic visual and audio content.
Tools like DALL-E 3 and Midjourney endow anyone with the ability to describe a scene in vivid detail with the creative ability of a seasoned graphic designer. (Albeit one that doesn’t have a great grasp on spelling.) Even if the finer details of these images are off and not quite up to the level of a six-figure display campaign, they represent low-lift, high-quality mock-ups that a designer can easily run with.
Basically, it makes it easier for less talented artists to communicate their ideal designs to those who are talented. Not to mention, it’s a major upgrade over stock images—just take a look for yourself:
Generative AI is also changing the game regarding video and audio content creation.
Tools like Synthesia give you access to hundreds of stock characters that come to life, speaking your exact copy. You can even create deep fakes of yourself or your peers using a composite approach with tools like D-ID and ElevenLabs.
Ross did just this a few months back. It’s eerie and clearly not him, but we might not be far off from a time when AI Ross is a lot more convincing.
Written First Drafts Are a Thing of the Past
This is one of the generative AI trends I have quite a bit of personal experience with—not to mention qualms about.
After ardently avoiding tools like ChatGPT for the first few months of 2023, I relented and started using it here and there to help speed up the content creation process. At first, I purely used it for initial topic research, text summary, and generating alt text: all tasks for which large language models like ChatGPT are relatively well-suited. But it turns out to be quite the slippery slope.
With the ability to attach plugins for link reading, peer-reviewed research, and YouTube transcriptions, those initial tasks turned into creative outlines and rough (I mean rough) drafts. The fact that these AIs spit out copy that was hollow and repetitive or over-the-top corny didn’t take away from the fact that they spit out passable content at a rate that’s orders of magnitude faster than me.
I feel mildly comforted by the fact that I’m not the only one thinking about the impact that these generative AIs are having on the earlier stages of content creation:
In an era where brands are aiming for that elusive medium between content quality and quantity, using generative AI for a first draft becomes a necessity.
Don’t get me wrong; I will still open the occasional blank Google Doc or even crack open a notebook to get things rolling, particularly for higher-priority pieces. But I have definitely come to accept that being a centaur also means being a lot different from the copywriters of old.
It’s Not All Sunshine and Rainbows
While this piece is dedicated to covering AI trends in a more positive light, it’s important to note that there are some serious (and valid) concerns as well. From deep fakes taking misinformation to a new level to a new category of jobs outsourced to machines, there’s a reason pioneers of artificial intelligence are scared, and governments across the globe are sitting down with industry leaders. A general-purpose technology with this much power needs to be contained where possible.
It’s important to look past the flashy headlines and astronomical revenue projections to consider the real human impacts of this technology. This technology can easily harm people, whether it’s through more elaborate scams or the evaporation of jobs.
So, as we move into another year of getting to know our AI teammates, it’s important to keep a balanced view of the good and the bad.
