The arrival of financial to the high positions of the companies brought an important demand for marketing to focus their efforts in the most efficient way possible.Artificial intelligence in marketing became extension in the next logical response after algorithms and content customization.
Artificial intelligence in marketing
The use of customization and constant improvement algorithms have existed in marketing many years ago. En sus inicios se trató de esfuerzos estadísticos simples como el A/B Testing que no es más que un modelo de inferencia bayesiana que busca que las evidencias u observaciones determinen la probabilidad de que una hipótesis pueda ser cierta.With time and the advance in the computing power of the great technological players by the hand of a greater volume of data, we pass from statistics to the experimentation of artificial intelligence.This word has been used so much that it seems like food additive and in marketing is as common as "sweetened with stevia".Today most of Marshas or Marketing Technology presume to have artificial intelligence systems that make their service superior.
However, it is enough to read Kai-Fu Lee in AI 2041 to understand that there are hundreds of different levels of this technology.From Deep Learning (set of algorithms that model high -level abstractions in data using computational architectures) and Big Data, to adversarial generative Networks (a type of artificial intelligence algorithms that are used in non -supervised learning, implemented by a two system of twoNeural networks that compete in a kind of zero sum game).With reading two definitions, we realize that they are not issues of popular consumption and that we rarely understand their true implications.In marketing we have limited ourselves to artificial intelligence doing the most efficient things.
This approach is very dangerous, artificial intelligence can make a creative process more efficient but less successful.A super simplified example of excessive power that we give to these tools is Solid Gold Bomb that devised an automated process to create t -shirts based on the British slogan "Keep Calm".The automated system created an iteration with "Keep Calm and Hit Her" among many others.In this example, the computation process was simple, but caused irreparable damage to the brand that in 2013 closed because of the scandal.This example happens every day, when writing "Tlaxcala" on Google you will see in the first options the recommendation "Tlaxcala does not exist" or write "Chuck Norris" will see another list of interesting results.
The effects of artificial intelligence on social media marketing
Artificial intelligence is also present in recommendation engines, Facebook spent a good part of 2021 defending the effects of its artificial intelligence motor on the impact they have on adolescents before the American Congress.Instagram recommendations have controversy effects on gender identity behavior, an extremely interesting reading of Financial Times explains very well (see here).In 2019, Instagram presumed the use of artificial intelligence in the generation of user recommendations (https: // ai.Facebook.com/Blog/Powered-by-Ai-Instagrams-Explore-Reommender-System/) In its article the Instagram team details how they build the recommendations from artificial intelligence applied to their huge audience-more than one billion active users-.The article ends with “the scale of both the community and the Instagram inventory requires enabling a culture of high speed and efficiency experimentation of developers to reliably recommend the best of Instagram for the individual interests of each person”.The key phrase is "culture of experimentation".A year later Instagram was in front of Congress defending itself, admit it or not, the artificial intelligence engine is at least part of the problem.
In marketing we must reinforce that human value remains our greatest contribution, the supervision and fine touch of a campaign builds results.Certainly the technology will make us more efficient but in the process it will have - for the child - to be wrong to “learn”.We must be careful not to make the technological tool omnipotent.Certainly artificial intelligence is key;For example, it detects fraudulent operations, but also preventing a person from doing a position with their card - without an explanation.
The purpose will be to establish a midpoint in which we take advantage of the benefits without taking the risks.