AI company Anthropic is facing criticism for its cooperation with religious groups, political ties, and exposure to misinformation, raising fundamental questions about AI ethical standards.
Imagine if the artificial intelligence (AI) assistant you converse with and ask questions of every day were actually thinking based on the advice of people with specific religious or political beliefs. If the direction of the answers provided by technology we believed to be neutral could change depending on what values it learns, the influence would be significant. A series of recent news items surrounding Anthropic, one of the leaders in the AI industry, is raising this heavy topic of “AI morality.”
Why is this important?
AI is not just a calculation tool. It now answers our daily questions, provides policy advice, and even plays a role in determining the direction of social discourse. If a company is closely connected to specific organizations in an non-transparent way when setting AI’s ethical standards, there is a risk that the information provided by the AI could reflect the interests or beliefs of a particular group. We must always be vigilant about whether AI is viewing the world fairly or if it is under someone’s influence.
Easy to Understand: Who determines AI’s values?
The moral standards that AI learns are often called an “AI ethics framework” (a code of conduct that AI must follow). Simply put, it is creating the rules that an AI must follow when conversing. Think of it like this: just as a photo app has beautification filters, AI has filters that screen its answers. Anthropic has revealed that it received advice from Christian religious leaders and theologians during the design of these filters Source: Anthropic Consults Christian Religious Leaders on Ethical ….
This is not just technical optimization, but a philosophical question about “what is right.” It is similar to teaching a child about the world based on the perspectives of a specific community, not just their school teachers. If this process were transparent and achieved through social consensus, there would be no problem, but if influence is exerted through closed channels, controversy over fairness is inevitable.
Current Situation: Shadows Behind Spectacular Growth
Anthropic is enjoying spectacular economic growth. Its revenue in the second quarter was approximately $11.5 billion, a 14-fold increase from the previous year Source: Квартальная выручка Anthropic взлетела в 14 раз до $11,5 млрд, and there are projections that its corporate value could reach $2 trillion when it goes public in October Source: Anthropic может стоить $2 триллиона при выходе на биржу.
However, controversy follows growth. Since the activation of its web search feature, instances have been discovered where Anthropic’s chatbot ‘Claude’ spread false information by citing Russian state-controlled media Source: Anthropic’s AI Chatbot Is Leaning More on Russian and Iranian …. Furthermore, as a member of the technology network TechNet, Anthropic has also engaged in political activities, such as opposing certain AI regulation bills Source: Firms crack down on AI propaganda operations, Anthropic and …. Recently, it also agreed to pay a $1.5 billion settlement for copyright infringement, and legal disputes continue Source: Anthropic AI copyright infringement lawsuit isn’t a total win ….
Where do we stand?
The AI industry is currently at a critical crossroads where it must control not only technical performance but also the impact that technology has on society. If the values of a specific group are deeply involved in the AI’s judgment criteria, there is a risk that the AI will prioritize the beliefs of that specific group over universal ethics. This is not just a problem for Anthropic, but a common challenge faced by all artificial intelligence used in our society. We must constantly ask the question: “Whose perspective” is the answer provided by AI?
What will happen in the future?
Going forward, companies will face the major task of proving the “moral transparency” of their AI. Beyond simply building a smart AI, they must be able to explain to the world what values the AI bases its judgments on. Anthropic calls itself a “public benefit corporation” and says it is committed to maximizing the benefits of technology and reducing its risks Source: Home - Anthropic. However, the public and oversight bodies will need to continuously verify how thoroughly these declarations are followed in actual operations.
References
- Anthropic Consults Christian Religious Leaders on Ethical Frameworks for Artificial Intelligence Development
- Firms crack down on AI propaganda operations, Anthropic and others
- Anthropic’s AI Chatbot Is Leaning More on Russian and Iranian Propaganda
- Anthropic AI copyright infringement lawsuit isn’t a total win
- Home - Anthropic
- Квартальная выручка Anthropic взлетела в 14 раз до $11,5 млрд
- Anthropic может стоить $2 триллиона при выходе на биржу
- IT Technology Alliance
- Christian religious leaders and theologians
- Journalists' Association
- Because its responses were too slow
- Because it spread false information by citing Russian state-controlled media
- Because it only provided paid services
- Data center security issues
- Copyright infringement lawsuit
- Unpaid employee wages