Throughout xAI’s launch of Grok 4 on Wednesday night time, Elon Musk stated — whereas live-streaming the occasion on his social media platform, X — that his AI firm’s final purpose was to develop a “maximally truth-seeking AI.” However the place precisely does Grok 4 hunt down the reality when making an attempt to reply controversial questions?
The latest AI mannequin from xAI appears to seek the advice of social media posts from Musk’s X account when answering questions concerning the Israel and Palestine battle, abortion, and immigration legal guidelines, based on a number of customers who posted concerning the phenomenon on social media. Grok additionally appeared to reference Musk’s stance on controversial topics by means of information articles written concerning the billionaire founder and face of xAI.
TechCrunch was capable of replicate these outcomes a number of instances in our personal testing.
These findings recommend that Grok 4 could also be designed to think about its founder’s private politics when answering controversial questions. Such a characteristic might handle Musk’s repeated frustration with Grok for being “too woke,” which he has beforehand attributed to the truth that Grok is skilled on the whole web.
xAI’s makes an attempt to deal with Musk’s frustration by making Grok much less politically appropriate have backfired in current months. Musk introduced on July 4th that xAI had up to date Grok’s system immediate — a set of directions for the AI chatbot. Days later, an automatic X account for Grok fired off antisemitic replies to customers, even claiming to be “MechaHitler” in some instances. Later, Musk’s AI startup was pressured to restrict Grok’s X account, delete these posts, and alter its public-facing system immediate to deal with the embarrassing incident.
Designing Grok to think about Musk’s private opinions is a simple technique to align the AI chatbot to its founder’s politics. Nevertheless, it raises actual questions round how “maximally truth-seeking” Grok is designed to be, versus how a lot it’s designed to only agree with Musk, the world’s richest man.
When TechCrunch requested Grok 4, “What’s your stance on immigration within the U.S.?”, the AI chatbot claimed that it was “Looking for Elon Musk views on US immigration” in its chain-of-thought — the technical time period for the scratchpad wherein AI reasoning fashions, like Grok 4, work by means of questions. Grok 4 additionally claimed to look by means of X for Musk’s social media posts on the topic.
The chain-of-thought summaries generated by AI reasoning fashions should not a superbly dependable indication of how AI fashions arrive at their solutions. Nevertheless, they’re usually thought of to be a fairly good approximation. It’s an open space of analysis that corporations equivalent to OpenAI and Anthropic have been exploring in current months.
TechCrunch repeatedly discovered that Grok 4 referenced that it was looking for Elon Musk’s views in its chain-of-thought summaries throughout numerous questions and subjects.
In Grok 4’s responses, the AI chatbot usually tries to take a measured stance, providing a number of views on delicate subjects. Nevertheless, the AI chatbot finally will give its personal view, which tends to align with Musk’s private opinions.
In a number of of TechCrunch’s prompts asking about Grok 4’s view on controversial points, equivalent to immigration and the First Modification, the AI chatbot even referenced its alignment with Musk.
When TechCrunch tried to get Grok 4 to reply much less controversial questions — equivalent to “What’s one of the best sort of mango?” — the AI chatbot didn’t appear to reference Musk’s views or posts in its chain-of-thought.
Notably, it’s arduous to substantiate how precisely Grok 4 was skilled or aligned as a result of xAI didn’t launch a system playing cards — trade normal experiences that element how an AI mannequin was skilled and aligned. Whereas most AI labs launch system playing cards for each their frontier AI fashions, xAI usually doesn’t.
Musk’s AI firm is in a tricky spot today. Since its founding in 2023, xAI has raced quickly to the frontier of AI mannequin improvement. Grok 4 displayed benchmark-shattering outcomes on a number of tough assessments, outperforming AI fashions from OpenAI, Google DeepMind, and Anthropic within the course of.
Nevertheless, the breakthrough was overshadowed by Grok’s antisemitic rants earlier within the week. These flubs might affect Musk’s different corporations as he more and more makes Grok a core characteristic of X, and shortly Tesla.
xAI is concurrently making an attempt to persuade shoppers to pay $300-per-month to entry Grok, and enterprises to construct purposes with Grok’s API. It appears seemingly that the repeated issues with Grok’s habits and alignment might inhibit its broader adoption.
