Corporate Influence as the Modern ‘Thought Police’ of AI Research – Elephant in the Lab

Scientific inquiry has long been at the heart of societal progress and innovation. Yet today, it faces subtle but profound pressures that are reshaping the questions scientists can ask and the projects they can pursue. Increasingly, funding patterns and institutional dependencies are shifting power away from public institutions and towards actors whose priorities may not align with open intellectual exploration.The era of curiosity-driven research, once protected by tenureship, is increasingly giving way to a system that favours politically safe and commercially viable outcomes, following market pressures, rather than genuine intellectual exploration. A recent investigation by The Guardian illustrates this dynamic starkly, where Sheffield Hallam University reportedly halted a major human-rights research project after major pressure linked to Chinese authorities. This raises concerns about the growing vulnerability of academic institutions to political influence (Hawkins, 2025). Recently, the Financial Times also reported that British universities’ reliance on tuition fees from Chinese students is fostering self-censorship among academics and administrators alike, who fear that sensitive research topics could jeopardise critical revenue streams (Borrett and Hughes, 2025). Together, these examples point to a broader trend: the gradual erosion of curiosity-driven research in favour of work that is politically safe or commercially advantageous. This shift narrows the range of questions that can be openly pursued and represents a quiet but significant challenge to the independence of scientific inquiry.

For better or for worse, we must acknowledge that this is not a new phenomenon. The shift in research priorities and power dynamics has unfolded across three distinct spheres: industry, public research institutions, and universities, each affected in different ways. In many OECD countries, publicly funded academic and institutional research has steadily lost ground, while industry-led research, especially applied and commercially oriented, has grown rapidly.  In the United States alone, federally funded research and development (R&D) funding fell from 1.86% of GDP in 1964 to 0.63% in 2022, while the private sector’s share rose to over 70% of total national R&D funding (NCSES, 2025). In Europe, the same trend follows: OECD countries report public R&D budgets have stagnated while the private sector funding has increased enormously in a multitude of industries (OECD, n.d.). This pressure is compounded by the growing demand that universities demonstrate societal impact, often through metrics and funding criteria that favour applied research and immediate utility over long-term foundational knowledge. As a result, universities and public research institutes, once the primary sites of foundational, curiosity-driven research, now operate in an environment increasingly shaped by commercial incentives and strategic industrial priorities.

The shift from public to private funding is particularly pronounced in the field of artificial intelligence (AI). This logic is often articulated through US–China competition, where AI leadership is framed as a marker of national power, legitimising the concentration of research capacity within private firms. It is especially evident in the ways in which the race toward artificial general intelligence (AGI) has become a proxy for both economic and geopolitical dominance. As private firms consolidate power over AI research, questions of ethics, transparency, and accountability have been sideline, and governments have allowed this, citing the convenient AI arms’ race narrative (Brennan, Kak & West, 2025). In an Orwellian twist, the pursuit of knowledge has been recoded as the pursuit of profit—what Philip Mirowski calls the shift from “truth-seeking” to “market-seeking” (2011).

AI Research Under the Corporate Lens

Public funding often aligns with public interest and includes funding for basic research driven by curiosity rather than immediate commercial value, corporate funding is driven by financial goals. This ‘soft money’ translates into real power: AI research and development is guided by corporate interests, mimicking a modern version of an Orwellian Thought Police, where science is guided by incentives rather than decrees. Dissenting thought may not be punished directly, but concurrent thought is grossly rewarded through salaries, career advancement, and publication prestige. 

While total investment in AI has soared to unfathomable heights, its distribution is extremely skewed. In both the United States and the European Union, the majority of new AI R&D spending now comes from large corporations, which overwhelmingly direct funding toward the development of Large Language Models (LLMs) rather than toward broader or more socially oriented AI research. In 2024, nearly USD34 billion went to LLM companies, while companies working on applied AI areas such as healthcare, education, or robotics received far less (Axis Intelligence, 2025; Mukherjee, 2024). Public funding mirrors this: as federal R&D budgets stagnate or decrease, agencies increasingly co-fund or align their grants with private sector priorities, especially where there are commercial or geopolitical returns, such as in the case of AI (OECD, n.d.). Basic, theoretical or cross-disciplinary AI research is sidelined in favour of applied, commercially profitable research. 

This concentration of resources in areas of AI research that corporations have deemed lucrative results in a self-perpetuating cycle: corporate labs have the computing and data infrastructures required for cutting-edge research; researchers from universities are attracted to high funding budgets to turn theoretical research into practical applications. The catch is that, more often than not, these researchers work on projects that feed corporate interest, rather than critical, high-risk, and possibly non-lucrative research. Meanwhile, universities struggle to fund computing capacity and public agencies increasingly justify grants based on impact metrics or short-term deliverables, aligning their evaluation criteria with market logic.

From Public Mission to Corporate Alignment 

The same public funds that were once used in  the pursuit of socially critical knowledge, often without commercial payoff, are now being redirected by industry partnerships or policy priorities shaped in dialogue with experts, often hailing from big corporations. Currently, governments are justifying alignment with the private sector through the narrative of an AGI ‘arms race.’ This is the idea that Western democracies must remain at the forefront of AI innovation and accelerate it at an unprecedented pace to stay ahead of geopolitical rivals such as China and maintain strategic advantage (e.g., proponents of an AI arms race have compared it to a “Manhattan Project” for AI and warned that competition with rival states will intensify if one side falls behind). [Big]Tech corporations often invoke this framing to redirect public funds towards initiatives that are aligned with corporate interests, securing both public investment and political legitimacy for commercially oriented AI R&D (Romero, 2025; Schmid, et al., 2025). 

This convergence of research interests leads to national and regional-level AI strategies being increasingly defined by public-private partnerships, creating new dependencies between regulations and companies that they are supposed to regulate. The effect is significant: corporate objectives shape research priorities and the metrics that define success. 

This alignment with corporate interest risks the loss of entire areas of scientific inquiry due to lack of monetisation or measure. Researching environmental sustainability applications of AI remains sidelined, even as evidence by the environmental cost of training LLMs becomes more substantial (de Vries, 2023). LLMs dominate investment not only because they are commercially viable, but also because corporations have promoted the narrative that they are the primary path to AGI, while alternative approaches are dismissed as inefficient or unpromising (Hao, 2025).

A similar example follows in the natural sciences, where funding for how AI and life sciences intersect is skewed towards drug discovery and development, aligning with pharmaceutical interests. Companies like Isomorphic Labs and Insilico Medicine leverage AI for drug discovery, while large partnerships like AstraZeneca’s $555 million collaboration with Algen Biotechnologies focus on AI-driven therapeutic research (Mukherjee 2024; Axis Intelligence, 2025). In contrast, equally crucial applications such as using AI to predict protein-environment interactions or model protein behaviour in ecological contexts receive far less attention or funding (NSF, 2025). These patterns illustrate how corporate-aligned incentives are subtly reshaping research agendas in the natural sciences, privileging commercially viable projects over foundational or socially critical scientific inquiry.

Corporate Influence as the Thought Police 

Corporate influence extends to how knowledge is shared. Unlike traditional academic practice of sharing information to inform scientific inquiry and innovation, corporations have an interest in selective transparency in different forms such as: restricting access to datasets, publishing results only in press releases or patent filings rather than open-access journals, or requiring non-disclosure agreements for collaboration with university researchers. In addition, corporations increasingly influence the scholarly publishing system itself, sponsoring journals, conferences, or special issues in ways that privilege certain methodologies, research topics, or institutional affiliations. The result is a scientific culture in which corporate actors incentivise specific research agendas while simultaneously acting as gatekeepers. They determine the lines of inquiry that are feasible, who can participate, and also how the knowledge circulates. 

This is Orwell’s ‘Thought Police’ reimagined in a digital age: control not through prohibition, but through design. Researchers are guided to think within the architecture of corporate possibility. If science is meant to serve public interest, then funding and governance structures must prioritise independent research and diversity of thought. This means increasing investment into public research and reforming evaluation systems so they are not only privileging metrics based on market logic. It also requires transparency in public-private partnerships. 

Corporations have an important role in shaping AI research but their shaping of the global AI scientific research agenda showcases the need for a cultural shift to resist this quiet normalisation of scientific inquiry being shaped by corporate interests. Without this, the research landscape becomes increasingly uniform; the incentives that once nurtured scientific autonomy are now calibrated towards impact visibility, rather than epistemic value.