On the one hand the institutions, which with the AI Act have regulated the artificial intelligence sector at a European level, on the other the companies, which continue to make generative Artificial Intelligence applications and software available to users. The risk? That of wasting time and creating a two-speed motorway, which confuses an already complex and continually evolving panorama. Despite the availability of apps such as ChatGpt and Copilot for months, the general public has approached generative AI thanks to smartphones, especially the Galaxy S24 series presented by Samsung in January. In the last few hours, the target audience of Galaxy AI has increased further, with the arrival of generative artificial intelligence features also on the "old" Galaxy S23, Galaxy Z Fold 5 and Galaxy Z Flip.
Translated into numbers, according to estimates by Counterpoint Research analysts, it means just under 40 million devices in the world capable of translating phone calls and texts with AI, writing messages supported by a chatbot, even editing photos, exchanging people and objects, for new and, theoretically, misleading results. "Given the expansion, greater awareness is needed on the part of people. AI is now everywhere", observes Alessandro Piva, director of the Artificial Intelligence Observatory of the Polytechnic of Milan.
"Also because - he continues - applications of this kind, intended for the masses, are completely legitimate and not in conflict with the AI Act of the European Union which is at the forefront on the topic. The rule provides four levels of risk to keep in mind in the large-scale adoption of technology, and the vast majority of systems currently in use fall into the 'minimal risk' category, which includes free-to-use video games and apps."
It doesn't mean letting your guard down, especially when the truthfulness of a content is at stake. Galaxy AI lets you change the way you use many apps, including calling and texting. With "live translation", for example, it is possible to translate an audio phone call in real time, displaying the text of an interlocutor on the screen, already in your own language. Similar operation for chat and text file translation. The possibility of changing the meaning of a photo, moving subjects from one side to the other, even inserting new ones that were previously absent, could push the creation of fakes. Samsung, in the metadata of the individual images, inserts an indication of the content generated by Galaxy AI but this is not necessarily enough.
"In Italy we have a Guarantor who is very attentive to these issues, perhaps more than elsewhere. Following the European Union's approach on the AI Act, which required months of approvals and debates, any doubts due to the opening of a platform of AI to the general public will not go unnoticed - Piva continues - and will be the subject of evaluation and analysis.
The challenge of timing remains: it is quicker to use a technology than to regulate it but there is much more proactivity today than in the past. We are currently discussing copyright and the generation of digital content with trade associations and institutions. Let's expect regulation in this sense too, which balances creativity and transparency."
Reproduction reserved © Copyright ANSA