A Trump presidency may mean trouble for ChatGPT & other 'woke' AI models, thanks to Elon Musk

Elon Musk recently threw artificial intelligence programs, including ChatGPT, into the heart of political controversy, voicing his view that AI models have become “too woke” and “politically correct.”

Speaking at the Future Investment Initiative in Riyadh, Musk suggested that AI models, particularly those developed in the San Francisco Bay Area, inherit a “woke” mindset from their creators, shaping outputs in a way he deems concerning. Musk’s remarks add fuel to a simmering debate over political bias in AI and could set the stage for increased scrutiny, particularly if Trump returns to office.

Musk, who has been a very vocal supporter of Donald Trump, and has often been seen at his rallies, heads xAI — a competitor to OpenAI, Google, and Meta. Naturally, Musk has a vested interest in positioning his AI as distinct from competitors. Industry watchers suggest Musk’s critical stance could align with the Trump administration’s agenda, potentially giving Musk’s xAI a competitive edge if established tech giants face political roadblocks.

Elon Musk was one of the first people to back OpenAI in 2015 and is often recognised as one of its co-founders. Musk invested in OpenAI with the aim of developing AI in a way that benefitted humanity, remaining open and accessible. It was structured as a non-profit to ensure it stayed aligned with its mission.

However, in 2018, Musk left OpenAI’s board due to concerns about conflicts with Tesla’s own growing AI initiatives, especially in autonomous vehicles. His departure marked the beginning of an uneasy relationship with OpenAI, which worsened after it shifted to a “capped-profit” model in 2019. Musk felt this move contradicted the original vision, suspecting that OpenAI was becoming more commercially driven.

The organisation’s partnership with Microsoft further fuelled his unease. Microsoft invested heavily in OpenAI, gaining access to its technology and enhancing its own products. Musk has repeatedly voiced concerns that OpenAI’s close ties to Microsoft and its commercial model could lead to the development of powerful AI under a corporate influence rather than for open, global benefit. He criticised OpenAI for becoming less transparent and overly ambitious with advanced AI capabilities. Musk later founded his own AI company, xAI, intending to pursue safe, transparent AI development and to provide an alternative that stayed true to OpenAI’s initial goals.

Analysts have speculated that Musk’s relationship with Trump’s campaign may even land him a spot in a future Trump administration, where he could influence tech policy to favour his AI model over others.

Previously, the Trump administration targeted Big Tech for perceived political bias, including moves against platforms like Twitter and Google. Trump’s chosen vice-presidential candidate, JD Vance, has echoed these sentiments, calling out Google as particularly influential and dangerous.

Analysts believe that, if re-elected, Trump could revive these efforts, shifting focus toward AI models that display any perceived political tilt. One tactic could include stricter regulations around AI outputs or even restrictions on models exhibiting a “liberal bias.”

Several recent studies have shown that AI models indeed carry a range of political leanings, largely due to the diverse internet data on which they’re trained. One study by researchers at Carnegie Mellon University, the University of Washington, and Xi’an Jiaotong University revealed noticeable political biases across different AI models. Another analysis from the Hong Kong University of Science and Technology found that open-source AI models often display liberal biases, especially on polarising issues like immigration and climate change.

As Musk sees it, these biases point to the “woke mind virus” infecting AI, a phrase he has used to criticise what he views as extreme political correctness in tech.

Experts argue that subtle biases are embedded in AI models through both the data they process and additional filters designed to prevent offensive content. The more these models interact with real-world data, the more these biases may appear, causing concern among some researchers.

Yet, Musk’s xAI model Grok, which Musk claims to be “maximally truth-seeking,” also faces similar issues, leaning conservatively on certain topics. With his assertive critiques and competitive positioning, Musk is betting that his platform will capture a growing audience concerned with political correctness in AI.

As the US election approaches, talk of “anti-woke” AI may only get louder, especially if Trump reclaims the Oval Office, adding an entirely new dynamic to the AI landscape.