By Tuhina Pandey
‘I write like a human’ is the new differentiator in the content world. While this may sound like a crafted response to the ever-growing dominance of generative AI (GenAI) which is disrupting the content space, it’s worth evaluating the possibilities and limitations of the GenAI in a creator economy.
Your imagination is your top skill
Think of a world where imagination is your key skill and the technical language needed to code, produce or create content is a simple language prompt. The applications of GenAI from creating content to delivering a hyper-personalised customer experience is immense.
GenAI is redefining creativity itself, what was creative yesterday is well within the realm of an intelligent machine to deliver – writing a blog, making a website, image, music and videos. The possibilities are endless, built to scale and ready to be delivered 24/7 in an automated manner.
What is more critical is to ask the right questions of GenAI with a specific purpose or task. We on the other hand should sharpen our focus on domains that are still beyond GenAI, things like the human imagination, abstracts, complex systems, critical thinking, irony and humour.
Good content needs to connect with the audience in a given context and if you are in the content business authenticity is non-negotiable. You must ensure that it does not draw or replicate from someone else’s work without guardrails and giving credit to the sources.
This puts a significant roadblock in the adoption of GenAI by businesses. GenAI may not be intended to plagiarise but lends itself to it given its dependence on aggregate and secondary data. As much as we enjoy weaving a poem or a song with AI, in a business environment plagiarism is a clear concern among other limitations.
GenAI – how far and no further?
On the whole, from the current vantage point, it’s clear that GenAI has transformative capabilities, and it will not only change the content game but disrupt it. I will even go further in saying that it will have long-term impact on consumer behaviour, expectations and experience. However, it is the transition phase that is crucial and in the short run, there is much to reflect on getting the foundation right.
So, what are the pitfalls of GenAI in the content world? And how can we navigate them?
- The data dilemma:
The jury is still out on whether the sea of secondary data when meshed with a prompt can generate something new or primary. And, by the same argument isn’t human creativity also an expression of assimilated experiences and therefore secondary data?
Well, realistically if you anchor your thinking on the need for accuracy and relevance of outcome – you rather work with data that is primary and proprietary to your business needs. Data that will stand up to the scrutiny of transparency, biases, authenticity, copyright laws, privacy and governance frameworks.
So, in a business environment it’s hard to deploy GenAI models that feed generic secondary data laced with these challenges. It makes more sense to use foundation models which are fine-tuned to your specific needs using data that is authentic and verifiable. Once you have a foundation model, the journey becomes cost and resource efficient as you can derive and draw from it for multiple use cases.
There is no free lunch, period. You need investments in high-quality data input, transparent training, foundation models, security, governance and upgrades for generative AI to grow and evolve in your business context.
- Perils of Hallucination:
Imagine if you ask GenAI models questions that are beyond their syllabus. In other words, what if it is beyond their training and therefore, application abilities? It will churn things for sure but not necessarily accurate or factual. From products to plans, you may end up getting seemingly right answers but in effect, it’s GenAI’s imagination gone wild and can distract and mislead you from the purpose of it.
- GenAI is vulnerable:
If AI is handled incorrectly, it can certainly lead to crisis. Imagine in the world of public relations if its application goes awry, it can lead to a crisis that damages a brand’s reputation. In extreme cases, if it’s used to spread misinformation at scale, it can be hazardous for society.
It’s critical therefore that we don’t miss the woods for the trees. Even if it takes time and diligence, it’s worth it.
- Human Sensibilities
This is clearly one of the most limiting aspects of GenAI. Intelligent beings and sentient beings are distinct. I am not a soothsayer but in the foreseeable future, AI may simulate feelings but does not feel it or for that matter has self-awareness. It is still just a tool, be it with tremendous potential.
It’s important we understand GenAI’s scope and limit our expectations and leverage it best for what it ought to do – generate.
Our focus should be to go beyond the romantic notions of GenAI to real business applications. From song sheet to balance sheet GenAI has ground to cover and that’s the task at hand for us.
The author is CMO and director of IBM marketing and communications, India and South Asia