Apple wants to “open the field of possibilities” on AI, but will do so in an opaque manner


Corentin Béchade

February 29, 2024 at 7:40 a.m.

0

Apple's future AI could well infiltrate Siri © Wachiwit / Shutterstcok

Apple’s future AI could well infiltrate Siri © Wachiwit / Shutterstcok

The iPhone maker will have taken his time, but now promises to invest in generative AI. Unfortunately, this will come at the cost of the transparency that should accompany this type of project.

Weeks of rumors and finally a confirmation. During a shareholder meeting, Tim Cook confirmed that Apple is indeed entering the race for generative AI to tickle ChatGPT, Bing Chat and other Gemini. Taken by surprise by the shift towards conversational agents for many months, Apple is therefore working hard to arm its next operating systems with features relating to AI, even if it means canceling ambitious projects like the Apple Car .

“Unlock new opportunities”

The CEO of the firm even explained that he wanted “open the field of possibilities» on the domain, convinced that this “will unlock new opportunities for users“. If the manager otherwise remained quite discreet about the concrete projects that concern Apple at the moment, he still had a word for the Vision Pro which could benefit from machine learning functionality to continue to improve the experience.

Unfortunately, this work of optimizing artificial intelligence will be carried out in the greatest secrecy. A practice not so foreign to Apple, but which could pose a problem here. An investment fund holding shares in the company has indeed proposed that Apple unveil the “ethical guidelines» followed by its artificial intelligence models, citing the “many social responsibility issues» that such innovations might otherwise trigger.

The need for transparency

The project was quickly dismissed by other investors (along with others concerning equal employment policies and the correction of gender and ethnic pay gaps), despite the support of Michael Forsythe , a major union figure within the company. On the side of Apple workers, there is fear that AI will destroy jobs, while on the consumer side, the proposal highlights the potential risks of violation of privacy.

Recent mishaps also highlight the need to be more transparent about the development of AI. The controversy created by Google’s AI, which distorts historical realities by creating images of black German soldiers in 1943 or representations of Native Americans at the table of the American founding fathers, proves that efforts to correct bias are still far away to be sufficient. Developing an AI model in complete secrecy and with a small group of people (and therefore necessarily less representative of the diversity of society) is not likely to help things.

The 5 best artificial intelligence chatbots (2024)
To discover
The 5 best artificial intelligence chatbots (2024)

Sep 15, 2023 at 09:00

Service comparisons

Source : Bloomberg



Source link -99