As media consumption continues to grow exponentially, keeping track of news and trends has become increasingly challenging for businesses and organizations. In response to this, media intelligence and monitoring professionals have been working to provide valuable insights and analysis to their clients. However, with the emergence of new technologies like ChatGPT, there may be a more efficient and effective way to navigate this space. To explore this possibility further, we spoke to three media intelligence experts about their thoughts on ChatGPT's potential impact on the industry.
​
(Disclaimer – ChatGPT wrote this opening segment to showcase the language model's capabilities).
​
The language models are getting more advanced, which, at first glance, opens many new opportunities for media intelligence companies. ChatGPT, in particular, could help with content creation, translations, summarisation, and even sentiment analyses. However, Carlos Alfredo Diaz, the CEO of GlobalNews Group, takes this technology with a grain of salt: "ChatGPT is a chatbot, built to showcase large language models (of the GPT3-3.5 family), it's main feature is to prove it can understand user prompts and provide cohesive responses (although not necessarily factual ones)." With that in mind, Diaz doesn't believe this particular AI gadget has any true potential in our industry, unlike language models in general.Â
​
With that being said, ChatGPT and similar technologies are a double-edged sword: "One of the positive aspects can be cost savings due to lower personnel," says Fady El-Murr, the Founder and Managing Director of Pressrelations. However, he can imagine this innovation will bring new competitors and players onto the market, as well as fewer clients who need us. The danger of replacement seems possible even to Simon Ernst-Sunne, the CEO of Opoint: "It definitely has the potential to threaten our consulting and advisor business in Media Intelligence, as we are viewing it today." But not everybody is so skeptical. Diaz, for example, does not think ChatGPT will destroy our business. According to him, the main threat lies somewhere else – the technology is perceived as something that it is not, and it does not perform efficiently.
​
That brings up another important dilemma – can we present language models as a bulletproof tool? After all, ChatGPT was trained on data, which is now outdated, and if asked who won the latest Super Bowl, the answer is that in 2021, The Tampa Bay Buccaneers defeated the Kansas City Chiefs (while in reality, the Chiefs actually defeated the Eagles this year). Carlos Alfredo Diaz explains that we have to keep reminding ourselves of the limitations: "ChatGPT is a conversational AI which has had some supervised learning in order to offset bias, but it is not a truthful-knowledge AI, its knowledge is merely statistically significant and is not intended to be more than a showcase of its conversational capabilities." Additionally, Fady El-Murr looks even further when he talks about his fear of society being at risk of blindly consuming biased information. He says that once humans trust a machine, they are open to being manipulated: "The manipulation will be based on the ethical frames of its creators or of those who control it. Ask yourself if you trust google maps or if you trust your own skills in using maps?"Â
Eventually, the question is not what if – the transformation has been happening, and we need to learn how to deal with the consequences. But it might not be as bad as some fear. Ernst-Sunne assumes language models will keep making our jobs easier. Moreover, there will always be a place for advisory and consultation: "I would argue that it is actually going to be
more rewarding working with media intelligence because we start at a higher level of knowledge," he says. On top of that, El-Murr adds that our advantage is that we will still be data collectors, and it is crucial to keep this upper hand.
Â
Without a crystal ball, it is hard to predict which particular forecasts will come true. But it is clear that ChatGPT and similar language models have the potential to transform the media monitoring and intelligence industry. If you still doubt it, create an account and test its limits yourself. And who knows, maybe you'll find yourself in a similar situation as Fady El-Murr, who admitted he caught himself talking with the machine for hours about ethical constraints.
Comments