X, a subsidiary of Elon Musk, recently updated the FAQ page of its AI chatbot Grok, heralding that this controversial feature will be launched soon. Musk revealed this mysterious model back in April last year, saying Grok would provide some "offensive, inappropriate and offensive" responses under the model, figuratively likened to "a still learning." amateur comedian”. The launch of this model will undoubtedly further challenge the boundaries of AI ethics and user acceptance.
At present, "Out of Control Mode" has not been officially launched. TechCrunch tried to find the feature in Grok's web interface, but it did not succeed. This new model may be a key step for Musk to achieve Grok's original vision. When Musk first launched Grok two years ago, he emphasized that it would be a marginalized, unrestrained and opposed "awakening culture" chatbot that can answer controversial questions other AI systems don't want to touch. This vision may be more thoroughly reflected in the "out of control mode".
Among the existing features, Grok has shown its unique qualities. When users ask them to use vulgar language, Grok is able to fluently output some swear words that are hard to hear in ChatGPT, showing a "unrestrained" style. However, when it comes to political topics, Grok still seems to be reserved, especially on sensitive issues such as gender identity, equality projects and diversity. Studies have found that Grok's tendency is more biased towards the political left. This phenomenon has caused widespread discussion and controversy.
Musk blamed Grok's performance on his training data and promised to make him "more politically neutral." "Unfortunately, the internet is full of nonsense about awakening culture. Grok will get better, it's just a beta," he said in a social media post in December 2023. Ske hopes to adjust the training data to allow Grok to handle politically sensitive topics more balancedly in future versions.
Additionally, Musk and many of his supporters, including Trump’s crypto and AI adviser David Sachs, accused existing AI chatbots of censoring conservative views. Sax specifically mentioned OpenAI's ChatGPT, saying it was "programmed to awaken" and was not realistic enough on politically sensitive topics. This criticism reflects the complexity and controversy of the political stance of the AI field, and also provides more context and motivation for Grok's "out of control mode".
Overall, Grok's "out of control mode" has not yet been launched, but its arrival may bring more controversial and offensive responses to users, further promoting Musk's vision for this AI system. The launch of this model will not only test users' acceptance of AI ethics, but may also trigger a new round of discussions about the role of AI in political and social issues. In any case, Grok's "out of control mode" will be a bold and controversial attempt in the field of AI.