This week, if you purchased an AI tool, you may have made a significant political decision, whether you were aware of it or not. Here’s why this is crucial and what steps we can take moving forward.
The Pentagon issued a stark ultimatum to Anthropic: either allow unrestricted military utilization of your AI models, or forfeit your federal contract. Dario Amodei, CEO of Anthropic, stood firm against this demand. He specifically opposed two applications: widespread domestic surveillance of U.S. citizens and fully autonomous weapons systems devoid of human oversight.
In retaliation, President Trump instructed all federal agencies to cease using Anthropic’s technology and set a six-month deadline for the military to eliminate it from their operations. Following this, Defense Secretary Pete Hegseth classified Anthropic as a national security “supply-chain risk,” a designation typically associated with foreign adversaries rather than domestic firms.
Shortly after, OpenAI revealed a new agreement with the Pentagon, allowing its models to be used on classified networks. While OpenAI asserts that it has maintained ethical safeguards, the Pentagon has not clarified why it accepted these limitations from OpenAI while banning Anthropic for similar reasons.
In light of this, I have decided to shift my subscription to Anthropic.
Founded in 2021 by researchers who left OpenAI over concerns regarding AI safety, Anthropic is unique in that it retains all of its original co-founders, demonstrating a robust culture in the face of commercial pressures. Amodei has asserted that AI can help defend individual rights and democracy, but that potential is not guaranteed. It requires concerted efforts from corporations, governments, and citizens alike. This reality is not merely a slogan; it’s fundamental to understanding the conflict we’ve just witnessed.
As I noted in a previous column, visible and intentional consumer choices can serve as a form of accountability. In Hawaiʻi, which is notably integrated into the online economy, we possess greater leverage than we often utilize. As a household, we consciously chose to support a company that aligns with our ethical values, even when many others do not.
Nonetheless, consumer pressure alone falls short.
Before the recent developments, Hawaiʻi was already lagging in AI governance. The 2026 legislative session offered valuable opportunities, but progress has been slow. The federal government will not establish necessary rules here. An administration that blacklisted a domestic firm for refusing to engage in mass surveillance is unlikely to enact meaningful AI regulations.
So, what steps should Hawaiʻi take? A useful model can be found in Pennsylvania, which conducted a pilot program with ChatGPT and published its findings. Government employees reported an average time savings of 95 minutes per day using the tool while only dedicating about 35 minutes daily to it. This indicates tangible benefits.
However, the report also highlighted that the AI fabricated nonexistent legal cases and created false job qualification requirements—errors that only human oversight could catch. Pennsylvania concluded that while AI serves as a beneficial tool, human involvement is essential to ensure accuracy and ethical compliance.
As we consider AI adoption, we must deliberate what our state policy will be. Are we intentionally selecting which AI vendors to engage with, and why? These are critical questions for both legislators and the public. We possess the power to determine how we utilize these technologies and which tools we support with public funds.
We should also prioritize data protection and look to Utah as a legislative example. Starting this July, the Digital Choice Act will mandate that social media platforms allow users to download and transfer all their data—including social connections, comments, and interaction history. Additionally, it will require platforms to enable interoperability, freeing users from being locked into closed systems. This approach is not simply a response to one problematic practice but a transformative change that grants users actual control, rather than relying on the benevolence of platforms.
Hawaiʻi requires similar protections, and we are currently behind many other states in implementing them.
AI regulation represents one of the few issues currently enjoying genuine bipartisan support. Various states are implementing measures that require chatbots to identify themselves as non-human, restrict AI use in insurance, and safeguard children from AI companion apps. These laws are being enacted in both conservative and liberal states. Utah, for instance, is not often viewed as a progressive legislature. There exists a demand for regulations; Hawaiʻi’s leaders simply need to take action.
The situation concerning Anthropic serves as a poignant reminder of what is at stake. This incident is not solely about one company’s ethical stance; it underscores the wider repercussions of unrestrained technology policies. A corporation chose not to create tools for domestic surveillance or autonomous weapons without human supervision, and the administration viewed this as insubordination. Such a development is concerning. However, we should remember that we are not powerless.
Consumers possess the ability to influence the trajectory of AI through their purchasing decisions. Concurrently, lawmakers have the responsibility to establish regulations that protect society from potential risks. For Hawaiʻi to genuinely shape its digital landscape, both forms of action are imperative and demand immediate attention.
