Ethical boundaries for usage of generative AI is key to brand success: Experts

ASCI in collaboration with Khaitan and Co released a white paper which acknowledges the applications of Generative AI in advertising and addresses concerns around the vast potential, legal considerations, and the usage of AI in creating automated ad content.

Speaking at a panel discussion on Generative AI, Asavari Moon, Global Marketing Leader, Google London, emphasised that the usage of AI in marketing is not new. Earlier it was used for getting recommendations be it in shopping carts, music streaming platforms or even OTT and digital platforms in the form of suggestions via predictive analytics.

“With Gen AI, there are three main categories where we will see the maximum impact- audience targeting (segmentation and personalisation), content creation and customer experience  (through chatbots or automations),” she stated pointing out the three big things where GenAI will make a difference in her viewpoint.

Amit Bhasin, Chief Legal Officer, Marico, said, “AI has certainly sparked a new debate in the world of advertising, especially for the folks in the FMCG sector, regarding how important it is to understand consumer behaviours and what works with a particular consumer be it catch lines or even key messages seen in an ad.”

All of this, in his view, is going to become easier with the passage of time to become more targeted and focused in terms of advertising that you can create for different consumers but with the same product.

“What kind of content is going to work with which target audience, is there a tagline, discount or offer that is going to work for a consumer are some of the different parameters that are there from a marketing point of view that companies will and are going to use as we move forward in this journey,” he added.

He further went on to add that while the debate has already started around the same in the advertising world, people are still evaluating the various use cases for trial and error since as corporates, everyone is trying to see what things we can really play along with, and how it can be put to the best use of the business.

Upon being questioned as to what is the general sentiment within the deployment of AI as a working tool, Ankit Bose, Head of AI, Nasscom AI, replied that while AI has been existing since 1940-1950s, the way it was operating was that certain sets of data were being fed into it and it was mainly supervising it until it began discriminating or differentiating between what is what.

“In the current Gen AI model, the transform model, what it does is that it self-learns, basis the data provided to it, develops a pattern and then generates new differentiated content, which could be in four formats- text, audio, images and video,” he said.

He then went on to add that from an adoption perspective, there are two broad categories when it comes to which companies are adopting and how they are adopting. One category of enterprises who are consuming or adopting is in favour of banning it all and then seeing how it goes, whereas the other is ready to go one exception down in terms of certain codes of ethics that they would want to try.

In his viewpoint, the challenge when dealing with AI or even Gen AI include business value, commitment from executives and choosing the right technology since one needs to be primitive with Gen AI as the biggest issue therein is a lack of skills since the technology basically generates something from what is fed into it.

“Currently, it is not in a position to replace humans because it will always need humans. Therefore rather than replacing humans what it tends to do is increase the productivity of the human. For example- If a marketer took probably seven days to create content, with GenAI it will be reduced to probably two days, giving the marketer a window to generate much more content and even better options for his clients or his internal consumers,” he stated.

Similarly, Manisha Kapoor, CEO and Secretary General, ASCI, also shared that in her viewpoint, GenAI is at a ‘much’ experimental stage as of now and that people who are curious about the technology and what it can do are actually going to derive the best answers.

“Technology today is really spending technologically and therefore with so many Beta Versions, at this stage, a lot of it is actually experimental where people are trying to see what all is possible in terms of learning versus what actually gets out to consumers, with a strong filter of what are the do's and don'ts and the guardrails that we need to build because the skills that you develop in this investing area that are going to determine how relevant you are in the next few years,” she said.

Commenting on the fear in ad and marketing professionals pertaining to job displacement, Moon went on to talk about the three main challenges that she had bucketed for easy reference.

“One has lots to do with risk, brand safety, and the adherence to brand voice because of legal concerns; second with investment and third being certain brand guidelines that marketers have in terms of tonality, the voice of the brand, etc. because when one prompts GenAi to generate outputs and if they’re not trained well, then one does not get the desired quality or consistency, or even an output that is completely off-brand,” she said.

That being said, she also suggested some of the ways to come around these challenges and suggested that since GenAI is ‘Garbage in, Garbage out’, it is essential to ensure that the input or data that one adds in is right because basis that would the output be right and hence fine-tuning the AI model using one’s own data to ensure that the output meets the desirable quality standard is crucial.

“It's also marketers' job to review and edit that content to ensure that it's consistent and inclusive and at the same time doesn’t have a lot of bias. Secondly, there should be a legal consent right from ethical boundaries to infringement of copyrights or intellectual property rights by establishing some very clear guidelines and standards for AI-generated content,” she said.

Given that AI technology needs a substantial amount of investment as well, Moon stated that since there is a need for getting such new technologies and tools and use them in daily life, marketers should consider giving that support, particularly around resources and upskilling to their employees to ensure that there is a smooth transition.

Additionally, Bhasin also pointed out that unknown infringement on someone’s IP, issues of data privacy or even breach, alignment to brand ethos, etc. are some of the challenges that come along with Generative AI.

In terms of adoption of the GenAI-led tools and technologies across the globe, there is a ‘phenomenal’ opportunity for India, at a country level, to understand the value chain of Gen-AI which has four levels- core chip manufacturer or the novel LLM manufacturers, applications developed on top, cloud providers and deployment or usage.

While there is no gap in the second and fourth levels, there is a gap in the first and third levels since India doesn’t have any of the core chip manufacturers or even cloud providers, he opined.

Upon being questioned as to what advice she would give to advertisers who are considering putting or incorporating Gen AI in their advertising strategies, Kapoor replied that since the idea really is to learn and be cautious about how it plays out in the public domain, there are issues to watch out for with regards to the emerging technology.

“My advice would be to experiment with it, learn from it, but just be careful in terms of how you do it,” she said.

Info@BestMediaInfo.com