A year ago, when ChatGPT appeared, it seemed that artificial intelligence (AI) was going to transform the world of work and the world at large. A year later, that revolution has not happened… yet. Although 40% of Spanish companies already use new programs derived from this technology, they automate simple administrative tasks, without producing the huge number of layoffs that dozens of surveys have projected. Experts predict that AI will take over entire sectors. It will happen in fashion, in the health sector, in the legal sector and in the administrative sector. When? Impossible to know. It could be six months, a year or five.
Meanwhile, there are fields in which AI is already present. One of them is talent acquisition. “We have been working with predictive algorithms for a long time, which allow us to predict the future searches that our users will carry out,” says Anna Via, project manager at machine learning of InfoJobs. The analysis of information from those who aspire to fill a job is the fuel for many selection algorithms. Thanks to these tools —and the information they are able to manage—, users of this employment portal can more easily locate offers that fit their profile.
“As we are able to know what your last work experience was, we can offer you vacancies related to it. We are also able to locate available positions based on the analysis of the interactions that other users with a similar profile have had,” explains Via. That is, an unemployed IT professional will find in their profile on the website offers similar to those of their last job or those that fit them due to their training.
A LinkedIn report revealed that four years ago, in 2020, 67% of hiring managers and recruiters were already using database-powered algorithms at some point in the selection process. Seven out of ten companies in this sector are trying to implement this practice, according to another study by Deloitte.
Tourism, and specifically hotels, are another testing ground already explored by AI. “The chatbots [programas capaces de mantener una conversación] They are being used a lot in the management and modification of reservations, because their skills to interact with customers have advanced a lot,” explains Álvaro Carrillo, general manager of Hotel Technology Institute (ITH).
However, it is in customer acquisition that the true potential of these tools is revealed. “They allow companies to identify who their potential customers are on the Internet, and where and how they should run advertising campaigns to attract them. This requires dealing with a lot of data that AI is capable of managing,” adds Carrillo. These programs, for example, adapt advertisements to the most representative style and patterns of the different social networks.
Efficiency and optimization
Chatting with robots is now a reality that is far from science fiction. “We employ virtual concierges,” admits Susana Garrido, director of learning and digital innovation at Les Roches and director of Spark Innovation Sphere, the innovation centre of this international school specialising in hotel management and tourism. “They provide information, for example, about the hotel’s meal service times,” she explains.
This technology has not only taken over customer-facing tasks, but has also been deployed within companies. “In the customer service department, housekeeping [limpieza de hogares] “We are experimenting with tools that allow for better management of staff and greater optimization of resources,” adds Garrido. That is, the program determines how many people are needed and for how long based on the rooms that need to be cleaned.
The infinite applications of AI, in fact, are changing companies of all sizes, which see in these new tools an opportunity to improve their productivity. Companies such as PotenzzIA“We have two types of clients: those who contact us without really knowing what they need, simply because of the current expectation; and those who really know what they are looking for,” explains Iván Prada, founder of this Valencian company dedicated to AI consulting.
There is only one previous step that blocks the entry of the less well-off: the need for a database to provide the necessary information. “Since 2000, what we are seeing is a process of lowering the cost of everything involved in implementing the final product, so it is not so much a question of money. It has more to do with the fact that in order for these programs to be implemented, there needs to be a certain digitalization of the company,” Prada adds.
Technical profiles
The high capacities available to algorithms are displacing the workers who were previously in charge of these tasks. That is why companies now demand profiles that not only know how to use these new tools, but also have sufficient criteria to supervise and evaluate their results. “In schools we are already seeing this transformation in the way our students work, and we are adapting our methodology so that they learn to use artificial intelligence and do not take everything it tells them as correct,” says Garrido.
“The way we understand work is going to change,” says Garrido (ITH). “But I think that, in our sector, something that we do very well in Spain will continue to take precedence: dealing with the client. And the best way to do this at the moment is a person.” A thought shared by Garrido, from Les Roches: “There will be clients who will look for hotels where there are only robots and others that do not. But the idea of someone coming to reception and being served by a humanoid robot is still a long way off,” he adds. “Even if there are algorithms that can carry out tasks better and faster than a person, at the point we are at, the presence of a human being within the chain is still essential,” Prada agrees.
Determining the timeframes for the evolutionary process of AI is impossible, even for those who work every day on the creation of these new tools. However, progress is unstoppable. “There is no horizon as such. We are in a constant evolution in which, suddenly, leaps occur,” says Arnau Vendrell, co-director general at iSocialWeb, Grower and in the artificial intelligence consultancy NeurekkaThese jumps, despite being erratic, are framed within predefined scales.
Five levels
The American company OpenAI, developer of ChatGPT, defines five levels of training of artificial intelligence: the first includes those that can carry out certain conversations (chatbots); the second, to reasoners, prepared to understand and solve problems at a human level; the third, to agents, systems that can take action on behalf of a user; the fourth, to innovators, who promote the improvement of the tools that exist in the environment in which they are developed; and the fifth, to artificial intelligences that by themselves can do the work of an entire organization (organizations). “If we use these five levels as a reference, we can say that we are approaching the second,” considers Vendrell.
The main problem with technological advances is that they are not deployed equally throughout the business fabric. “The current reality is that companies are in a somewhat strange situation,” says Guillermo Tato, a specialist in the adoption of artificial intelligence in the technology industry. Raona“The same thing happened when the Internet came out: everyone wanted to use it, but few knew exactly what for. Now there is a boom in AI, but it has not yet been possible to fully marry the companies that are in need of finding cases in which to apply these new solutions, and which have precisely that technological know-how.”
Aside from the aptitude scales, which define the generic skills that these tools will gradually acquire, Tato focuses on sectors and activities that will be revolutionized by AI in the short term. “I think it will have a great impact on the fashion sector thanks to the generation of images. It is something that is already there, but it has enormous potential. Soon it will be possible to put together an entire photo shoot of certain garments without the need for the model or the setting to exist,” he points out.
Raona, with 20 years of experience in the development of technological tools, has implemented Neural Fashion, a system that already explores some of these possibilities. It is as easy as entering the image of a garment and asking the program to generate an image with a model, male or female, in a given scenario.
Reliability
The administrative environment —since it is the most prone to storing information through databases— will be another playground for AI, capable of chewing through mountains of data without difficulty. “Something that is about to arrive are systems capable of classifying invoices, regardless of their format and where they come from. Although there are programs that already do this today, they have not been given the final capacity to execute payment or collection orders, because human supervision is still recommended. But we are already reaching very high levels of reliability in carrying out these processes,” Tato explains.
A similar reduction in tasks will also occur in the legal environment. “I think that law firms will soon have programs that allow them to create contracts, or review the clauses necessary to generate official documents, and that will be able to give consented responses to certain notifications,” Tato insists. However, in his opinion, if there is one environment where AI will have a special capacity to impact in the future, it is the healthcare environment. The day is not far off, he says, when programs capable of recording and transcribing medical consultations will arrive, while assisting the doctor in preparing the report using the same technical language. An aid that will facilitate the delivery of more personalized care and that will relieve the doctor of administrative tasks. “Until now, some experimentation has been done with this, and it was discovered that patients were much happier that the doctor could pay more attention to them, instead of being focused on the computer.”
Free up resources
The potential of AI to free professionals from tedious tasks is particularly significant for freelancers and smaller companies, who see in these tools the possibility of improving their productivity. However, lack of knowledge – both conceptual and practical – remains an obstacle for many. “We have a challenge as a country with training and we have realised that when you present this technology and its capabilities to small businesses, their initial reaction is one of surprise,” admits Pedro Andueza, president of ATA Canarias, and vice-coordinator of the CEOE Data Economy and AI working group.
“In Spain, in general, we have a low level of digitalisation in our productive fabric,” says the representative of the employers’ association. Something that plays, at the same time, in favour and against: “Some of these applications, such as ChatGPT, do not require great knowledge, quite the opposite, and can allow companies unable to make a costly investment to be more versatile and free up resources to dedicate to the growth of their business model.”
Future prospects frighten some and give hope to others. “People believe that AI will make our lives worse, but they say that out of fear,” agrees Vendrell, from Neurekka. “Whatever role it ends up having, it is clear that it will be in our lives,” adds Tato, from Raona. “Faced with some ethical questions regarding its use, I allow myself this reflection: AI is, simply, the result of what we are as people. And its use, as with many other things, will depend on the hands in which it is placed,” concludes Andueza, from ATA.
Follow all the information of Economy and Business in Facebook and Xor in our weekly newsletter
#work
