Grok (chatbot)
Grok is a generative artificial intelligence chatbot developed by xAI. Based on a large language model (LLM), it was developed as an initiative by Elon Musk in a direct response to the meteoric rise of ChatGPT, the developer of which, OpenAI, Musk co-founded.[5] The chatbot is advertised as "having a sense of humor" and direct access to X.[6][7] It is currently under beta testing and is available with X Premium.
This article is about the Grok chatbot. For the word, see Grok. For other uses, see Grok (disambiguation).Developer(s)
United States, Canada, Australia, New Zealand, United Kingdom, India and 42 other countries[3][4]
Background[edit]
Musk co-founded the AI research organization OpenAI with Sam Altman in 2015. Musk left the company's board in 2018, saying of his decision that he "didn't agree with some of what OpenAI team wanted to do".[8] OpenAI went on to launch ChatGPT in 2022, and GPT-4 in March 2023. That month, Elon Musk was one of the individuals to sign an open letter from the Future of Life Institute calling for a six-month pause in the development of any AI software more powerful than GPT-4.[9]
In April 2023, Elon Musk said in an interview on Tucker Carlson Tonight that he intended to develop an AI chatbot called "TruthGPT", which he described as "a maximum truth-seeking AI that tries to understand the nature of the universe".[8] He expressed concern to Carlson that ChatGPT was being "trained to be politically correct".[10]
TruthGPT would later become known as "Grok", a verb coined by Robert A. Heinlein in his 1961 science-fiction novel Stranger in a Strange Land to describe a form of understanding.[11]
History[edit]
In November 2023, xAI began previewing Grok as a chatbot to selected people,[12] with participation in the early access program being limited to paid X Premium users.[13] It was announced that once the bot was out of early beta, it would only be available to higher tier X Premium+ subscribers.[14] At the time of the preview, xAI described the chatbot as "a very early beta product – the best we could do with 2 months of training" that could "improve rapidly with each passing week".[15]
In December 2023 the Silicon Valley start-up Curio launched a range of AI-powered children's toys, including a rocket-shaped character named Grok. The toy is voiced by Musk's ex-girlfriend Grimes, who is also an investor in the start-up, but the product is unrelated to the xAI service.[16]
On March 11, 2024, Musk posted on X that the language model would go open source within a week and six days later, on March 17, Grok became partially open source. Disclosed were the networks architecture and its weight parameters.[17]
On March 17, 2024, Grok-1 was open sourced under the Apache-2.0 license.[18][19][20]
On March 26, 2024, Musk announced that Grok would be enabled for all premium subscribers, not just those on the higher-end tier, Premium+.[21]
On March 29, 2024, Grok-1.5 was announced, with "improved reasoning capabilities" and a context length of 128,000 tokens.[22]
On April 4, 2024, an update to X's "Explore" page included summaries of breaking news stories written by Grok, a task previously assigned to a human curation team.[23]
On April 12, 2024, Grok-1.5 Vision (Grok-1.5V) was announced. Grok-1.5V is able to process a wide variety of visual information, including documents, diagrams, graphs, screenshots, and photographs.[24]
On May 4, 2024, Grok became available in the United Kingdom,[25] being the only country in Europe to support Grok at the moment due to the impending Artificial Intelligence Act rules in the European Union. Grok was later reviewed by the EU and was released on May 16, 2024.
Features[edit]
Humor[edit]
An xAI statement described the chatbot as having been designed to "answer questions with a bit of wit" and as having "a rebellious streak". It said that bot had been "modeled after The Hitchhiker's Guide to the Galaxy, so intended to answer almost anything".[15]
An extract shared by an X employee showed Grok being asked to answer the question "When is it appropriate to listen to Christmas music?" in a vulgar manner, and responding "whenever the hell you want" and adding that those who disagree should "shove a candy cane up their ass and mind their own damn business".[26][27]
The chatbot defaults to its "fun mode", which has an "edgy" and editorialized voice, but can be set to "regular mode" where it does not.[28] Elizabeth Lopatto of The Verge compared Grok's fun mode responses to the game Cards Against Humanity, and noted that while the chatbot could be aggressive in tone she had never seen it turn that aggression on the user.[29]
Political stance[edit]
The chatbot has been characterized as "anti-woke" by the press.[30][31] In response to Sam Altman, the CEO of ChatGPT developer OpenAI, Musk said "the danger of training AI to be woke – in other words, lie – is deadly".[11]
The chatbot has been marketed as having a willingness to "answer spicy questions that are rejected by most other AI systems".[15] Elon Musk shared a screenshot of Grok giving detailed instructions on how to manufacture cocaine.[32][33] Musk noted that Grok's responses were limited to information already publicly available on the web, which could also be found with regular browser searching.[34]
Following the chatbot's December 2023 launch to Premium+ subscribers, Grok was found to be giving progressive answers on questions about social justice, climate change, and transgender identities.[35] After research scientist David Rozado claimed to have applied the Political Compass test to Grok and found its responses to be left-wing and libertarian – even slightly more so than ChatGPT – Musk responded saying that xAI would be taking "immediate action to shift Grok closer to politically neutral".[36]
Accuracy[edit]
Testing Grok in December 2023, Vice's Jules Roscoe found that it was returning misinformation and false timelines when asked about news events, such as wrongly claiming that the Israel–Hamas war had reached a ceasefire in early October when it had not.[28]
When reviewing the chatbot's "fun" mode, Vice magazine found that the setting resulted in "both sides" responses when asked about debunked conspiracy theories such as Pizzagate, where its regular mode would identify the theories as false.[28]
Since April 2024, Grok has been used to generate summaries of breaking news stories on X. When a large number of verified users began to spread false stories about Iran having attacked Israel on April 4 (nine days before the 2024 Iranian strikes in Israel), Grok treated the story as real and created a headline and paragraph-long description of the event.[23] Days later it reacted to many users joking about the solar eclipse with the summarized headline "Sun's Odd Behavior: Experts Baffled".[37]