Google DeepMind's establishment of a full-time philosopher position has sparked widespread global discussion. According to public information, Cambridge University scholar Henry Shevlin is set to join the DeepMind team under the official title of "Philosopher". This is not an isolated case. Anthropic, another AI giant, has long integrated researchers with philosophy backgrounds into its model value-alignment work, the most notable being Amanda Askell, who is responsible for the design of Claude's value system.

Many view Google's hiring of a philosopher as a signal of a "renaissance in the humanities". However, the involvement of philosophers in AI laboratories does not mean that the traditional humanities are once again becoming prominent in the job market. What AI companies truly need is not humanities talent in the general sense, but individuals capable of participating in value judgments, rule design, and the construction of governance frameworks.

On the surface, Google's hiring of a philosopher appears to address the issue of AI value alignment. Yet, this in itself demonstrates that the development of AI is expanding beyond the scope of mere technical competition, increasingly involving value judgments and social consequences. What truly deserves attention is the public governance gap exposed by this shift.

In the past, tech companies mostly solved problems of technical feasibility, where the main competition lay in algorithms, computing power, and efficiency. However, as large models enter domains such as education, healthcare, finance, law, and public information dissemination, companies are increasingly confronted with value judgments regarding what should or should not be done, behavioral consequences, and the balancing of interests. How to balance freedom of speech with content safety, how to respond to users' emotional dependence, and to what extent AI should be allowed to influence human behavior are the questions that cannot be answered by increasing computing power or optimizing model parameters.

This is precisely the common background behind Anthropic establishing a persona-alignment team, OpenAI continuously discussing the social impacts of AI, and DeepMind hiring a philosopher. When AI begins to participate in value judgments and social interactions, engineering logic reaches its own boundaries. A model can answer questions, but it cannot define what the right questions are. An algorithm can execute rules, but it cannot determine whether the rules themselves are reasonable.

When technology begins to affect human cognition, behavior, and social relations, the role of the enterprise has actually shifted. Within traditional commercial logic, a company is a market entity with clear boundaries whose primary responsibility is to provide products and services. However, with the development of AI, tech companies are no longer merely providing tools. Instead, they are profoundly influencing the operation of society. From information acquisition to value judgments, and from labor relations to public discussion, companies' product decisions are producing increasingly evident public consequences. Enterprises have acquired growing social influence, yet they have not received the corresponding public authorization, accountability mechanisms, or governance constraints.

In this sense, the true governance gap at this stage is that the pace of technological expansion far outpaces that of the adjustment within governance systems. This means that market entities are effectively deciding an increasing number of issues with public impact, while existing governance systems have yet to establish corresponding checks and balances.

Recent discussions surrounding AI regulation in the United States serve as a real-world reflection of this governance gap. Although the federal government, state governments, and tech companies are all attempting to participate in AI rulemaking, the general picture remains fragmented. Regulations at the state level continue to multiply, whereas systemic legislation at the federal level progresses slowly.

Nevertheless, technological development does not wait for institutions to complete their designs. While legislative bodies are still debating rules, large model products have already entered the daily lives of hundreds of millions of users. What information users encounter, how they interact with AI, and which content is prioritized are increasingly shaped by product design and algorithmic mechanisms.

In truth, rules are not absent. Rather, they are encoded into products in advance. Technical systems are shaping reality while governance systems are still chasing technology. This reality itself is far more important than whether legislation is moving fast or slow.

This is precisely the deeper issue underlying Google's hiring of a philosopher. Philosophers are entering labs because companies are beginning to realize that they are facing an increasing number of problems that originally belonged to the realm of public governance.

However, can hiring philosophers fill this governance gap for corporations? The answer is likely not so optimistic.

Philosophers can indeed help companies establish value frameworks. Whether through risk reviews, ethical assessments, or designing the behavioral boundaries of models, they are essentially adding a layer of value constraints to technical systems. As AI integrates more deeply into social life, such constraints are clearly necessary.

Crucially, philosophers address internal corporate value problems, whereas the governance gap itself is a social problem. The former is about making judgments, while the latter concerns who has the right to make the judgments. Though they appear similar on the surface, they actually belong to entirely different levels.

For instance, an AI company could easily hire the world's finest team of philosophers to co-author a set of value principles and determine how the model should answer questions accordingly. Yet a new question arises: why should this company be the one to decide? Why not society? Why not through a public process?

Hence, the issue is no longer a philosophical one, but one of public governance. It is also no longer a question of value judgment, but of legitimacy.

Philosophers can help companies contemplate what constitutes a reasonable choice, but they cannot answer why the enterprise possesses the right to make such a choice. This also means that while Google's hiring of a philosopher is a noteworthy signal, it is not a solution to the governance gap.

From a practical standpoint, many of the issues brought about by AI have already exceeded the scope of internal corporate governance. Whether it is content governance, algorithmic influence, employment disruption, or the restructuring of the information order, the subjects impacted are no longer just corporate users, but society as a whole.

For these types of issues, relying solely on internal corporate ethics teams, risk committees, or philosophical advisors makes it difficult to cultivate sufficient governance capacity. The reason is that corporations possess technical capability but lack full public authorization. Meanwhile, governments possess legitimacy but may not keep pace with the speed of technological iteration. Academic institutions, media, and civil society organizations can provide expert opinions and public oversight, but they lack direct enforcement power.

What truly needs to be established in the era of AI is not corporations bearing the responsibility of governance alone, but rather a multi-layered governance structure capable of coordinating technological innovation, market forces, and the public interest.

For tech companies, the most significant shift may be moving from a product logic to an infrastructure logic. In the past, internet enterprises often viewed themselves merely as providers of tools. However, when hundreds of millions of people begin to rely on the same AI system to acquire information, form judgments, and even make decisions, these systems acquire increasingly obvious characteristics of public infrastructure.

As corporations gradually face the potential public consequences of their product decisions, transparency, explainability, and accountability mechanisms will progressively be integrated into the core agenda of corporate governance structures. At the same time, common rules at the industry level are becoming increasingly vital. Data standards, safety protocols, model evaluation systems, and risk management mechanisms cannot be accomplished by a single enterprise independently. As AI steadily becomes a crucial infrastructure for economic and social operations, the importance of industry governance will continue to rise.

Final analysis conclusion:

The most profound significance of Google hiring a philosopher lies not in a renewed appreciation for the humanities and social sciences, but in the fact that tech companies are beginning to realize that the development of AI has expanded beyond the scope of pure engineering problems. What truly deserves attention is the public governance gap exposed by this shift. Philosophers can help enterprises comprehend problems and formulate value frameworks, but they cannot replace society in rulemaking, nor can they confer public legitimacy upon corporations. The truly scarce resource may no longer be just computing power, algorithms, and capital, but rather the governance capacity to establish coordination mechanisms among technological innovation, market forces, and the public interest.

_____________

Dr. Liu Lidan is a researcher fellow at ANBOUND, an independent think tank.