Why artificial intelligence will never be able to replace us
A few days ago OpenAI released version 4.0 of ChatGPT and I'm sure your friends or colleagues were talking about the wonders (or threats) that this new version offers. From academic writing to web development, it seems that artificial intelligence is permeating all sectors and will sooner rather than later doom professions such as copywriter, illustrator and even programmer?
It does not seem that any (halfway serious) media outlet has considered replacing researchers or editors with artificial intelligence because of its limitations. Professor Charlie Beckett of the London School of Economics and Political Science runs a programme in journalism and AI and recently told The Guardian that "for newsrooms that consider it unethical to publish fake news, it is very difficult to implement ChatGPT without spending a tremendous amount of human time editing and checking that the information is true". That is the real limit of artificial intelligence: it cannot emulate basic human capabilities for most jobs.
The real limit of artificial intelligence: it cannot emulate basic human capabilities for most jobs
Dall-E surprised us a short time ago with a la carte illustrations worthy of the best designers. Just by entering a phrase, the system is able to return a multitude of graphic resources with better or worse technique. But who would want to have pictures painted by an AI hanging in their home? Probably someone who doesn't appreciate art and who usually leaves the film that comes with the frames 'because it looks pretty'.
So what will DALL-E end up with? With low-value illustration work, and will be able to replace images previously sourced from banks, allowing for greater customisation of 'stock resources'. However, no publishing house will consider illustrating a children's book based on AI, nor will the poster for the next SEMINCI be created by an algorithm. The reason is simple: in this type of commissions, human intervention is essential as it provides the necessary knowledge of the social and cultural context of the commission, applies their personal experiences and personalises the result.
We don't chop code or draw prototypes: we develop solutions
A similar situation exists with software development. Twitter user @itsafiz, a senior developer, posted on 15 March this year that ChatGPT-4 had been able to interpret a sketch drawn in a notebook and transform it into the web page with the sections indicated on it. And it is true, the website is fully functional and reproduces the sections that Afiz drew, but beyond that it does not add any value. The tool will also generate a website in a few seconds from the instructions we give it in text, but what kind of website? Like the one generated by Afiz, we are talking about developments of very little value and quality, which will serve for those who are looking to have a portal quickly and with extremely limited functions.
Potential recipients of a ChatGPT-generated website could be the owners of your village bakery, who only want to show photos of their products, contact details and links to their social networks. They are a very unprofitable customer for companies in the sector, who would normally use a standard template to complete the job, and will now use Artificial Intelligence to reduce production times and human involvement in production, increasing profitability by reducing costs on less profitable jobs such as these.
However, any more complex development will always require human intervention for one simple reason: software developers and user interface and user experience designers do not code and prototype, they design solutions to customer problems. First of all, the analysis of an organisation's needs is the most important step in the process: the collection of requirements, the study of the environment and customer needs, etc. are basic. And they depend on a social, economic and cultural context in which human intervention is central.
Laura Ródenas explained in the podcast on design 'En Cómic Sans' how the Cambodian airline Bassaka Air had not taken into account for the naming of the company something as basic as the sonority in different languages and had become a joke for the Spanish-speaking community. Applied to naming, an AI can offer us thousands of names based on the parameters we indicate, but it will not tell us if those names have a counterproductive meaning in any language, if the sonority is appropriate or if it is registered in any country. Again, human intervention covers many more fields and is essential.
Returning to software, during the development process, the experience of the teams contributes indispensable value to the quality of the software: previous solutions to similar problems, adaptation of the systems to the specific needs of the company or the territory, knowledge of the vulnerabilities or limitations of existing technologies to decide whether or not to implement them, etc. None of this can be done by an AI. What it can do is improve development times, reduce costs and increase code quality through its use by programmers or designers. Consulting options on how to develop a complex algorithm or how to perform a certain action in a language through tools such as ChatGPT will allow us to save time in Google searches or reduce the learning curves of certain technologies, but always so that those specific answers offered by the AI are subsequently analysed and applied by a professional.
Companies that know how to use these tools to their advantage "will be able to have more customers at a lower cost"
This is basically what has happened so far with industrial production, which has been replacing the most tedious and repetitive manual jobs with robots or automated systems. This revolution is now spreading to more sectors thanks to artificial intelligence, but it will never replace the real core of developments.
In short, artificial intelligence has come to improve the lives of professionals in different sectors, but never to replace them. Alberto Gómez, creative director of Azote Studio, believes that AI will occupy a place that design studios or development companies have not occupied until now, as they will be used by those who "have never seen the importance of investing in design, just like those who do not understand the importance of buying fresh food, bread made by a baker or clothes that are not the result of exploitation in developing countries". In the meantime, companies that continue to provide real solutions to their customers and know how to use these tools to their advantage "will be able to have more customers at a lower cost".