Google announces that it will train its AI language models with public content

A Google update last weekend introduced several changes to its privacy policy . Previously, it stated that " user data would be used to train language models ," mentioning only Google Translate. The updated version changes this term to " AI models ," specifically mentioning Bard and Cloud AI alongside Translate.

Since their inception, or rather since they became popular with the popularity of ChatGPT, one of the many ethical issues with generative AI systems like ChatGPT and Bard is the way they extract and use data. Now Google admits that it's going to use public data to train Bard and so on.

Google will use public data to train its AI Bard

Often, this information is public. But that doesn't eliminate plagiarism and privacy issues, nor the possibility that AI might misinterpret what's been said and offer outdated and obsolete answers. Even Google has warned its employees to be careful when using chatbots like its own Bard, as they can make unwanted code suggestions.

There is also the question of whether this type of data mining is legal. For example, OpenAI, creator of ChatGPT; is facing lawsuits over accusations that it illegally collects personal information from Internet users and uses it to create its products.

OpenAI is also facing a lawsuit for copyright infringement and privacy violations related to allegations that it used copyrighted books without permission to train its AI systems. The company allegedly copied text from these titles illegally without obtaining the copyright holders' consent and without giving them credit or compensation.

Google would do this for many Internet posts. That causes a lot of uncertainty about how you're going to treat the data, especially if you come across bad data and regulate it so you don't take it into account when asked.

Source: TechSpot