a maximal truth-seeking ai (11. 764)

The social media platform X (formerly Twitter) has introduced a new feature for its chatbot, Grok—for premium subscribers—a text-to-image generator comparable to Bing’s service (which would rather infamous refuse the prompt “please make me a picture with the winner of 2020 US presidential election) or Facebook’s but apparently with the safety protocols turned off. Whilst users can find boilerplate guidelines and guardrails, presented in the first-person, proffering caution when it comes to making deceptive, provocative or plagiarised pictures, a cursory trial yielded some messages surely none would endorse. Though all tinged by that particular, cutting-corners AI patina that’s far from a watermark, a trial yielded far more offensive and topical content ready to be shared.