Vivaldi says no to AI… for now!

Among web browsers available on the market, a certain Vivaldi again and again renounces the invader. And in a blog post, the company explains why it says “no” to AI.

In 2023, THE big web trend was inevitably a certain ChatGPT, democratizing a new form of artificial intelligence in the eyes of the general public. The world of web browsers has not been spared from this trend, with many examples incorporating this technology in one way or another.

No AI at Vivaldi

However, even though ChatGPT technology is now exported on board our vehicles (and elsewhere), Vivaldi designers have no intention of integrating AI on board their web browser.

Thus, according to developer Julien Picalausa: “ LLMs (for Large Language Model) themselves are simply not suitable as interlocutors, as search engines, and can only help generate a language with a significant risk of plagiarism. “.

Concerns about plagiarism… and false information

For Vivaldi, the fundamental problems caused by AI are two. Initially, the AI ​​may tend to resort to outright plagiarism.

Depending on the requests, the conversational agent may in particular reuse an original text, thus using material protected by copyright, which leads to plagiarism. Additionally, texts from social media and other user-provided sources may also contain sensitive and private information.

Added to this is another concern, with AI-generated content that is largely sourced from the Internet… with what that may include in terms of false information.

“ Early experiments with the technology resulted in chatbots that quickly began spouting offensive comments themselves, proving they were not fit for purpose. “, explains Vivaldi. The recent experience of DPD tends to confirm this.

So many elements which mean that the developers considered it wise not to integrate such a solution into Vivaldi. “ There is enough erroneous information circulating to risk adding more » concludes Julien Picalausa.

Source : Vivaldi