There is still no complete consensus on the definitions of AI and Gen AI and for many it is not entirely clear where machine learning ends and Gen AI begins. Meanwhile, the vendor landscape is seen as being the ‘wild west’.
Nonetheless, our panel of execs from across the media industry are highly engaged with the concept of Gen AI and see it as a crucial topic for the future of the industry.
Most companies are taking a proactive stance to understand the opportunities AI offers. Their employees are exploring and experimenting with AI, but there is a danger of being ‘thrown in the deep end’. There is a real need for company guidelines and industry regulations to help mitigate the risks, both operational and legal, in the use of AI, whilst still empowering execs.
In terms of content generation, the greatest impact of Gen AI may be less on premium productions, but more to enhance user-generated and long-tail content, reducing the barriers to producing high-quality programming.
There are clear perceptions that Gen AI can only be as good as the data it is trained on, that it may have a tendency towards the generic and the average, and that it can be a useful tool for creatives as opposed to taking the creative step itself.
There is some skepticism about the concept of Gen AI creating original creatives for advertising campaigns, with a view that AI may tend to ‘regress to the mean’ given that it is trained upon what has gone before. It may have more impact on the production of performance ads aimed at activation, than on ads promoting brand values.
There may be more initial opportunities for Gen AI to assist in the adaptation and personalization of campaigns to increase relevance and facilitate hyper-targeting. However, concern was expressed that excessive adaptation could actually damage long-term brand identity and equity.
For both advertising and content there is early evidence that Gen AI trained on data sets that reflect societal biases and stereotypes is in danger of reflecting those biases back, particularly when it comes to gender roles. Gen AI content may create a feedback loop or filter bubble.
There is a clear opportunity for Gen AI to use object recognition to recognize and categorize creative elements that can then feed into ad testing and campaign evaluation. This will have a major impact on market modeling and media planning.
Gen AI could also facilitate the democratization of media buying, lessening the barriers to entry for companies who may not have the skills to design ads and buy media.
Whilst Intellectual Property is a concern, particularly when it comes to the data used to train Gen AI models, there is pessimism about the ability to protect or monetize content in practice. With so much data in the public domain, paid licensing may not be a realistic option.
There is excitement around the use of AI to analyze data sets, offering both scale and speed to insights. Pattern recognition may lead to insights that would not otherwise be evident. Large language models and AI trained on past research and data may also offer opportunities to test campaigns and content without the need to do survey research, but again there is concern whether this would eventually lead towards generic rather than leftfield or game-changing creatives.
Ultimately at its best Gen AI can be an enabler, a tool that can augment rather than replace creativity, by automating labor-intensive tasks to allow execs to focus on what really matters, namely enhancing skills. At its worst, it could replace execs with an average version of themselves. In the long term, the threat to media jobs could come not from AI, but from those execs who have a better understanding, and make better use of, AI than they do.
