AI Made Friendly HERE

How generative AI can remove monotony from digital marketing, Marketing & Advertising News

<p>Representative Image</p>Representative Image
Much has been said about ChatGPT, and even more, seems to be happening around it. Companies and developers are quickly creating Plug-ins. Marketers and influencers are hastily creating ‘How to’ videos and posts. And even stand-up comedians are incorporating the Chatbot into their routines.

For anyone who has been living under a rock, ChatGPT is a ChatBot created by OpenAI that seemingly provides lifelike answers to input requests. This includes writing code, generating backlinks, drafting blogs, or explaining quantum physics. The attention it has been drawing is immense and the bot now has over 100 million users. OpenAI has now rolled out a subscription plan that is accessible only to those in the US.

The more I read and saw, the more curious I got. I’ve never seen any AI technology evolve so rapidly and garner so much attention. But along with it came feedback and information about how the bot was providing incorrect, biased, and sometimes harmful answers.

As the owner of a Digital Marketing Agency, I wondered what impact ChatGPT and Generative AI as a whole would have on the different functions within my organization and the larger ecosystem.

Generative AI became popular in the last decade with the advent of deep learning architecture such as GANS (Generative Adversarial Networks), VAE (Variational Autoencoders), and Transformer. These models were initially used to create realistic human images. The Transformer model, which could generate text and write software code, was a game changer as it was scalable.

With ChatGPT and DALL.E2, we have now come to a head where with better algorithms, larger datasets, and better models, these generative AI tools can create more realistic images and write long paragraphs of coherent text. I believe we have reached a tipping point, and it would be foolish to think that Generative AI will not have an impact on Digital Marketing.

Gartner expects that by 2025, 30% of outbound marketing messages from large organizations will be synthetically generated, up from less than 2% in 2022. I personally believe that this number will be higher. Organizations need to get a deep understanding of what these tools are capable of, what they can and cannot be used for, and how quickly they can adapt to the change while staying accountable for their output.

ChatGPT

We are not new to content writing tools. They have been around for several years. Simple tools such as Grammarly and Hemingway help content teams with spelling and grammar, while more advanced tools such as Copy.ai, Rytr, Frase, INK, Jasper, etc., generate copy for emails, blogs, Ads, websites, and social media posts.

There have always been concerns about using these tools. While the pros revolved around scalability (speed of churning out content), writing efficiency, and cost-effectiveness, the cons included plagiarism concerns, content quality, and the lack of creativity or originality. But the biggest risk for AI-generated content was Google’s algorithm update, which was rewritten to devalue the content written by AI tools. Google was clear that original content that adds value to the searcher was more important than writing for search engines.

However, ChatGPT has distinguished itself by providing one answer to your question or request. Think about consolidating all the responses on Google and providing one answer. It gleans through millions of information and provides an almost human-like response.

I asked my content team to experiment with ChatGPT. To feed it queries, commands, and requests and have a chat with it based on our service lines and the work we have done so far. All of them found the tool easy to use, and the results were mostly accurate. However, they also found some results to be very verbose, lacking in specifics (no dates, numbers, data, citations, or quotes), and repetitive despite reframing the question. It did put out incorrect information as well.

DALL.E2

Almost all designers will tell you that searching for the right images takes more time than creating the design itself. If Generative AI tools can provide custom-made images quickly, then designers can use the time they save for more creative thinking and high-end design work.

OpenAI’s ‘DALL.E2 Explained’ video on YouTube highlights how certain tasks (like replacing a dog with a cat without changing anything else in the background) can be done in seconds. It chooses the precise cat image (from among hundreds) which fits perfectly with the background in one-tenth the time it would take a designer to find the same image. Does it then make business sense for a designer to engage in this time-consuming, repetitive task? I would think not. Instead of going through reams and reams of stock images, it would be more productive for a designer to take the output DALL.E2 gives and refine it further.

My design team found that the tool could generate realistic images, but options were limited. Images of tech, such as crypto and cloud, were inadequate. Also, the images can only be downloaded as a .png file at the moment, which makes editing more difficult.

Will Generative AI Displace Human Workers?

The debate about technology replacing humans has been going on for years. A study by McKinsey in 2017 stated that in about 60% of occupations, around one-third of the tasks or activities could be automated, which means there would be a significant change for workplaces and workers. But there are always two sides to a story. A report by the World Economic Forum in 2020 states that by the end of 2025, technological progress (including AI) will create 12 million jobs.

On the surface, it looks like a threat. Questions like what would happen to content writing, writers, and designers can be heard even from my team. But I don’t believe it to be a threat. We have been using Jasper for over two years, and it has not replaced anyone. Companies that opt not to utilize content writers and instead rely only on tools to generate their content run the risk of publishing duplicate content. The content itself would be boring and impersonal while providing no value to the reader.

ChatGPT cannot solve everybody’s content problems. DALL.E2 may produce images faster, but it still requires finetuning, branding, and vetting to meet company standards. All my teams are united in saying that every output generated needed a once-over, a quick edit, some tweaking, or fact-checking.

It would be good to remember that AI tools are based on what came before. They can comprehend and generate, but they cannot create.

So how can you use ChatGPT and DALL.E2 in your organization?

As an organization, start by listing all your services. For each service, think about what ChatGPT or DALL.E2 can do and which services they cannot add value to. There might be services where they can help by completing almost 80-85% of the job. For example, a listicle blog written specifically for SEO purposes or a quick image for World Environment Day. For other services, they might be able to contribute only 5-10%. This could include value proposition documents or thought leadership articles.

With the feedback I received from my teams, I believe ChatGPT and DALL.E2 can be used to help with monotonous, repetitive tasks. The writer or designer can then utilize the time saved by adding more original and creative nuances to the piece or image.

My content team found the tool useful for quick research and answers, brainstorming ideas, and generating new ideas when they got stuck or faced writer’s block. High volume low impact work like SEO backlinks and Quora answers were quick to generate. Blog writers took the top 5 questions from answerthepublic.com for their topic and fed them into ChatGPT. They now had enough information to write a whole blog without the need to go through all the answers on Google. After adding data, insights, and anecdotes, the blog was personalized and ready.

The design team could generate images quickly where a combination of 2-3 elements was needed. They believed that the tool could be used for generic social media posts. Any time saved by the designer was then used to enhance the final output. The creative control still remained with the designer.

My thought leadership team asked ChatGPT for a weekly and monthly social media plan for the thought leaders they work with and found many new and useful ideas.

While it is tempting to see how we can use Generative AI to help our clients, start by using it to help your teams. How can it be used to benefit them? Where can you save them time so they can focus on more creative and intellectual work? How can Generative AI take the monotony out of creativity?

Using Generative AI Responsibly

The cons of Generative AI are evident. For every article admiring its capabilities, there are many that call out its dangers. My team agrees that some of the answers were opinionated, biased, and incorrect. Responsible AI is at the top of every organization’s list of things to do. From Google to Adobe, everyone is working on making AI more secure, safe, transparent, and fair.

In the meantime, as digital marketers, we need to acknowledge that while Generative AI tools might be responsible for the output, we are still accountable for what we publish and share. We must continue fact-checking, tweaking, refining, monitoring, and editing our work. In the long run, when we start using Generative AI for commercial purposes, we need to create standard operating procedures, policies, and fail-safes, and train employees on when and where Generative AI can and cannot be used.

Conclusion

Generative AI will change the way we think about creative work. It extends beyond Content and Design into Sales, Operations, and HR. If we do not accept it, it will be our loss.

In the end, using generative AI tools is like asking for advice. You listen to the advice and appreciate the input, but you also use your own judgment, experience, and common sense to decide what works best for you. It is like having your own personal assistant who can take away the tedious lowbrow work, leaving you to do more meaningful and creative work.

AI does not, at this point, have all the answers, and humans need to use their own intelligence and expertise to decide what works and what does not for their business. If Tony Stark can occasionally ask his most advanced and trusted AI machine Jarvis to shut up, so can we!

ChatGPT continues to amaze and startle a lot of people by the sheer genius and possibilities it holds, the author writes.

  • Published On Feb 16, 2023 at 07:48 AM IST

Join the community of 2M+ industry professionals

Subscribe to our newsletter to get latest insights & analysis.

Newsletter icon

Originally Appeared Here

You May Also Like

About the Author:

Early Bird