ChatGPT users delete app after OpenAI US Department of Defense partnership

A single announcement then transformed the atmosphere almost overnight.

Within days, OpenAI’s flagship AI assistant was hit by a surge of angry users, a wave of one-star ratings and widespread deletions after its partnership with the US Department of Defense was revealed.

A weekend that transformed sentiment around ChatGPT

OpenAI’s mobile app had long featured prominently in app-store rankings and was widely regarded as an innocuous productivity tool. That view began to unravel at the end of February 2026.

From 28 February, users began removing the ChatGPT app in far greater numbers. Figures reported in France indicated that deletions rose by around 295% over one weekend, an almost unprecedented increase for an established mainstream consumer app.

Users did not just walk away quietly. Ratings plunged, with bad reviews rising an estimated 775%, turning app stores into a venting space for frustration and distrust.

Previously, complaints about ChatGPT had largely concerned hallucinations, bias or subscription costs. This backlash has a different focus: rather than the product’s technical shortcomings, it centres on OpenAI’s ethical and political decisions.

The trigger: a deal with the US Department of Defense

The tipping point came with the announcement of a partnership between OpenAI and the US Department of Defense, which many overseas media outlets still call by its historic name, “the Department of War”.

OpenAI described the agreement as support for “defence and national security” use cases. That wording quickly alarmed users concerned about AI being deployed in warfare, surveillance or automated decision-making in conflict areas.

On social media, long-standing ChatGPT supporters described feeling betrayed. Many said they accepted AI being used for education, writing and coding, but considered any link with weapons systems or military intelligence unacceptable.

For a significant share of the audience, the partnership shattered the idea that ChatGPT was a neutral, civilian tool underpinned by purely benevolent intentions.

The dispute has also added to existing doubts about OpenAI chief executive Sam Altman, whose leadership has faced repeated scrutiny since the dramatic boardroom conflict in late 2023. Critics see the defence agreement as evidence that the company’s commercial and strategic priorities now outweigh its original safety-first message.

Bad reviews: what users are complaining about

From bugs and subscriptions to ethics and trust

Before this controversy, unfavourable reviews of ChatGPT’s mobile app commonly referred to free-tier restrictions, sign-in problems or GPT-4 access being limited to subscribers. Those concerns remain, but they have been pushed out of the spotlight.

Recent user comments have increasingly focused on:

• Anger over the perceived “militarisation” of a civilian AI tool
• Calls for greater transparency about government partnerships
• Concerns that data could be shared with defence or intelligence agencies
• Demands to boycott the app until ethical assurances are provided

Some reviewers explicitly say that, while they remain impressed by the technology, they will not continue to pay for it or keep it on their phones while these agreements are in force.

A crisis of trust, not simply a UX issue

The approximately 775% increase in poor ratings suggests users are doing more than expressing minor irritation. They are reconsidering whether the company’s objectives match their own values.

Apps commonly experience rating declines following software faults, redesigns or price increases. This reaction, however, followed a corporate announcement. That distinction is important: interface improvements and new features cannot necessarily mend an ethical divide.

Competitors see an opportunity

ChatGPT’s difficulties have predictably created advantages for rival services. Anthropic’s assistant Claude, for example, has publicly highlighted that it has not entered into an agreement with the US Department of Defense.

The company has referred to disagreements about potential AI uses in surveillance and autonomous weapons, presenting itself as more restrained on military applications.

By distancing itself from defence contracts, Claude presents an alternative for users who want advanced AI without ties to the arms sector.

Other organisations, including open-source projects and smaller start-ups, are likewise moving rapidly to promote their ethical commitments. Some point to transparent governance arrangements, limits on shareholders or explicit prohibitions on offensive military use.

Sam Altman faces renewed scrutiny

Sam Altman has always divided opinion in technology circles. Some admire him as a visionary who has brought AI into everyday life, while others accuse him of blurring safety research with forceful commercial growth.

The defence agreement has revived those arguments. Opponents contend that OpenAI has moved away from its original non-profit purpose and now operates like a typical contractor pursuing government work, especially in strategic fields such as defence.

Supporters respond that national-security work may involve non-lethal purposes, including cyber security, logistics, simulation and disaster-response planning. In their view, refusing all engagement with defence organisations would exclude AI from areas in which it could reduce harm or help stabilise crises.

Concern	Pro-deal argument	Critic argument
Use in warfare	Focus on defensive tools only	Tools can migrate to offensive uses
Data privacy	Contracts can include strict safeguards	Government access risk remains high
Public trust	National security partnerships are normal	Undermines civilian, friendly image of AI

What this means for everyday ChatGPT users

For most people, ChatGPT continues to operate in exactly the same way on their screens. It can write emails, create lesson plans and draft code. The algorithms behind those conversations have not suddenly become weapons.

The argument instead concerns governance and future direction. Users are effectively asking: if I depend on this tool every day, who will ultimately guide its future? And where are their ethical boundaries?

Some users are choosing to continue with ChatGPT while urging OpenAI to disclose further information about the partnership, including:

• Clear restrictions on the forms of military use
• Independent audits of safety procedures
• Public disclosures covering government contracts

Others are removing the app entirely and moving to alternatives with stricter stated positions on defence work, or to local models that operate on their own devices.

Understanding the risks and trade-offs

AI companies working with defence institutions face several overlapping risks. Dual-use technology may serve peaceful or violent purposes alike. A system trained to analyse satellite imagery for humanitarian-aid co-ordination could equally be used to support targeting systems.

There is also a danger of function creep. A model initially used for translation or logistics could later be fine-tuned for battlefield simulations. Even where a contract imposes limits at the outset, emergencies or political pressure may weaken those restrictions.

Conversely, a complete refusal to collaborate also carries consequences. Governments will pursue AI-enhanced capabilities regardless. If organisations with stronger safety concerns decline to participate, less scrupulous rivals may take their place and shape military AI with fewer ethical constraints.

How users can respond in practical ways

People troubled by the latest news have several practical choices beyond posting a one-star review.

• Compare the privacy policies and ethical commitments of different AI tools
• Use browser-based versions rather than mobile apps, and reduce data sharing where possible
• Try smaller or open-source models for sensitive work
• Contact providers to request explicit policies on military use
• Support research organisations and NGOs monitoring AI’s use in warfare

Another option is to divide tasks between services. Some users now retain a “general-purpose” assistant for harmless content while using local, offline tools for material involving personal, medical or political information. Separating these uses can reduce exposure if a provider’s partnerships change direction again.

For many people, the rise in poor ratings and app deletions is not so much a permanent split as a warning. It demonstrates that the public is watching the direction of generative AI, and that ethical limits set by users can prompt rapid, measurable backlash when they believe those limits have been crossed.