Elon Musk and X’s AI will have a special mode for offensive, inappropriate and insulting content.
On the FAQ-page from Grok a new ‘unhinged’ mode is mentioned. This option is described as an “amateur comedian still learning how to do it”. The mode itself is not yet active.
When Grok was unveiled by Elon Musk, he already said that his AI would be “anti woke” and would answer questions that other chatbots do not dare. For example, it is currently possible to use Grok to create political images featuring Donald Trump, while other AI companies stay away from this.
Data is running out
Elon Musk does have a problem, because the data to train AI has run out. During a live stream on X Musk said that all human knowledge has now been added and that AI now needs to be further trained with synthetic data: content generated by AI.
Companies such as OpenAI, Meta and Anthropic have long used synthetic data to train their AI models.
Read more about AI, or stay informed via the Bright Newsletter.
Source: www.bright.nl
