the attractive capabilities of AI smartphones – L’Express

the attractive capabilities of AI smartphones – LExpress

It’s a fish known throughout Silicon Valley. In his space opera as cult as it is crazy The Hitchhiker’s Guide to the Galaxy, British writer Douglas Adams invents the little Babel fish, a funny vertebrate capable of translating all the languages ​​of the universe when inserted into an ear. In 2024, this singular dream has become a reality – even if it is embodied in the banal form of the smartphone.

After mobile calls and devices connected to the Internet, the sector is entering a new era: that of AI smartphones. And their first representatives, like the Galaxy S24, which Samsung unveiled on January 17, already display attractive new features. These devices are notably capable of translating a conversation in real time, whether it takes place in a room or over the telephone. And the AI ​​doesn’t just offer an instant transcription: it generates an audio version, which allows the dialogue to remain natural and rhythmic, as if a human interpreter were at your side.

A translator in your pocket

“AI also brings a new way of working and getting information,” explains Bastien Sudron, smartphone product manager at Samsung France. Finished, the draw of lots for the unfortunate person obliged to take notes in the meeting and make a summary. Simply record a meeting and click on the Galaxy S24’s AI function to get a well-organized summary, in impeccable French.

A zealous assistant, the Galaxy S24 can also summarize any web page (news article, Wikipedia entry, etc.) in two seconds. A powerful productivity feature – AIs excel at synthesis. Samsung, however, communicates on this with a little too lightness. The manufacturer clearly indicates that the summary is generated by AI, but does not specify that it may therefore contain errors. However, 2023 is full of tragicomic examples of people having placed excessive trust in ChatGPT, like this New York lawyer who used it to develop a legal opinion, without realizing that the AI ​​was citing cases that it was inventing. all parts.

READ ALSO: France and AI: the underside of a formidable comeback

AI smartphones nevertheless offer undeniably attractive tools, particularly on a creative level. A cute cat Instagram post? A powerful presentation for a meeting? A photo of Seoul at night? It is no longer even necessary to capture, arrange or retouch anything: you just have to ask the AI ​​to “generate” them ex nihilo. Hordes of tourists armed with selfie sticks may soon be a distant memory.

“Many people – me first – spend time reformulating their messages. Here again, AI can save them a lot of time,” explains Bastien Sudron. From a raw frame, the artificial intelligence tool of the Galaxy range (Galaxy AI) offers the user several versions: one adapted to a professional setting, one with a more relaxed style, one with emojis …And, for good measure, a “polite” variation and another “#social” (designed for social networks, with hashtags corresponding to the subjects mentioned in the text).

READ ALSO: Generative AI: “There is nothing magical in it, only mathematics”

Samsung is not the only manufacturer to move its smartphones into the era of generative AI. Google also launched a new Pixel 8 range at the end of 2023 with functions boosted by artificial intelligence. Apple, which is actively working on its in-house large language model (LLM), Ajax, should launch several new functions on its iPhone and other devices in June. According to Counterpoint ResearchAI smartphones will represent 40% of the market by 2027.

Mini-AI in phones

The next harvests should certainly do better in two areas. First, the possibility of using AI without being connected to the Internet. The Galaxy S24 can only directly perform a limited number of tasks (translation, etc.). Many others, including image generation, are done in the cloud, which requires a connection. But the emergence of increasingly efficient mini-AIs should soon make them usable in all conditions. A dizzying prospect: even disconnected from the Internet, we would then have permanent access to a rich intelligence base – even if it is not devoid of imperfections, AI making mistakes.

We can also bet on the strong comeback of voice assistants, which will (finally) become practical to use. While Siri or Google Assistant only managed basic requests (“What’s the weather?” “Set an alarm for tomorrow”) and with annoying slowness, thanks to the immense progress made by OpenAI and its competitors, AIs are today able to understand surprisingly well what is said to them. The sweet dream of being able to chat freely with a machine rather than looking at its menus is gaining substance.

.

lep-life-health-03