By John Q. Hosedrinker 
Following initial backlash over concerns of loopholes for domestic surveillance, OpenAI has announced a reworked agreement with the Pentagon. The revised terms explicitly state that the AI system shall not be intentionally used for domestic surveillance of U.S. persons and nationals, and defense intelligence components are excluded from this contract. Despite these changes, some observers and legal experts remain skeptical, citing the lack of public release of the full contract and lingering concerns about broad interpretations of the terms. This development occurs amidst broader debates between AI companies and the military regarding ethical AI usage in national defense.
Read the original article here
It appears there’s a lot of concern and skepticism surrounding OpenAI’s recent adjustments to their deal with the Pentagon, particularly as worries about surveillance intensify. Many are questioning the sincerity of these changes, suggesting that they might be more of a damage control tactic than a genuine shift in principle. The idea that Sam Altman, OpenAI’s CEO, could so easily alter a Pentagon contract is met with disbelief, with some feeling he has already compromised his stance and is now merely trying to manage the fallout.
A significant point of contention is the lack of concrete evidence that any substantial contractual changes have actually been made with the U.S. Department of Defense. The narrative of massive subscription cancellations leading to a sudden realization of ethical wrongdoing is viewed with considerable doubt. Instead, there’s a sentiment that these adjustments are a reaction to a perceived decline in user trust and potentially financial pressure, especially with an IPO on the horizon. The suggestion is to maintain pressure through continued cancellations, arguing that even a temporary dip in revenue could significantly impact their valuation.
The core of the alarm stems from the potential for surveillance, and the specific wording in OpenAI’s updated agreements raises more questions than it answers. While the language now states the AI system shall not be *intentionally* used for domestic surveillance of U.S. persons and nationals, this implies that such surveillance might still be possible under certain interpretations or for non-U.S. persons. The exclusion of U.S. persons and nationals is seen as a loophole, and the question of whether U.S. citizens abroad would be subject to tracking is also a significant concern, highlighting a perceived willingness to spy on citizens prior to public outcry.
The situation also brings into question the performance and strategy of other AI companies, like xAI, which aren’t perceived to be facing the same level of scrutiny. This suggests a broader competitive landscape where ethical considerations are being weighed against commercial interests. There are also accusations of OpenAI actively trying to discredit competitors, such as Anthropic, by allegedly paying marketing firms to generate negative publicity. This perceived “hardcore damage control” fuels the distrust surrounding OpenAI’s motives.
Many individuals have already taken the step of deleting their OpenAI accounts and canceling subscriptions, expressing a firm stance against what they view as a betrayal of trust. The repeated declarations by Sam Altman are met with deep cynicism, with suggestions that his word cannot be taken at face value and that a physical verification of his statements, like checking the sky for blueness, is necessary. This level of distrust suggests that any assurances from OpenAI, particularly regarding government contracts, are likely to be met with extreme skepticism.
The fundamental issue is the perceived loss of integrity. Once a company’s ethical foundation is called into question, regaining that trust is seen as an almost impossible task, regardless of subsequent actions. The current situation is characterized as a desperate attempt to salvage a reputation, driven not by genuine remorse but by the threat of significant financial repercussions and a damaged public image. The narrative suggests that any “alterations” to deals are more about appearances than substantive ethical shifts.
The idea that OpenAI might be trying to maintain the original deal with the Pentagon but simply wants the optics to change is a prominent concern. The sentiment is that it’s too late for such measures; the public is aware of the potential for data exploitation once significant financial incentives are involved. The assertion that OpenAI and even the government are untrustworthy underscores a deep-seated skepticism that transcends specific companies and extends to broader institutions.
The feeling is that while power may have shifted towards consumers, the damage is already done, and many are not willing to return to OpenAI’s services. The analogy of being “fooled again” resonates, suggesting a repeated pattern of behavior that erodes confidence. The comparison to past bad deals, like Lando Calrissian’s negotiation with the Empire, highlights the perceived naivete of believing such agreements can be easily mended.
There’s also a view that OpenAI is falling behind in technological advancements while maintaining a premium price point, further diminishing their value proposition. The idea of them eventually forcing users to adopt their technology is seen as a threat that needs to be countered proactively. The contrast drawn between China’s overt approach to surveillance and control versus the perceived covert methods of Western companies like OpenAI is also a notable point of discussion.
The inability to undo past actions, like deleting an account, reflects a permanent break for many users who are now unwilling to re-engage with OpenAI. The specific language in the new agreement about not intentionally using the AI for domestic surveillance is seen as a direct contradiction to past statements, reinforcing the belief that OpenAI’s communications are untrustworthy. The call to action remains clear: continued pressure through subscription cancellations is essential, as OpenAI’s financial projections are heavily reliant on consumer revenue.
The notion that canceling subscriptions is ineffective is dismissed as either misinformation or a misunderstanding of OpenAI’s financial vulnerability. The article from Platformer’s Casey, referencing Keach Hagey’s insights into Sam Altman’s negotiating tactics, further solidifies the perception of deceit. Altman’s alleged playbook of saying whatever is necessary to achieve his goals, followed by undermining credibility if needed, paints a picture of a leader whose actions are consistently driven by self-interest rather than genuine principles. This pattern of behavior, characterized by dishonesty and chaos, is seen as particularly concerning for someone leading a company with such potentially world-altering technology.
The sequence of events, including the Pentagon’s actions against Anthropic and OpenAI’s subsequent deal, is viewed as a continuation of a trend where ethical considerations are often secondary to strategic partnerships. The concern that “Project CARNIVORE never went away, it just grew more teeth” is a chilling reminder of historical surveillance programs and suggests a continuity of government interest in data acquisition. The perceived deletion of news stories further fuels suspicions about transparency and a coordinated effort to control the narrative.
Ultimately, the prevailing sentiment is one of deep distrust and disillusionment. OpenAI’s attempts to alter their deal with the Pentagon are seen as a desperate reaction to a crisis of confidence, driven by financial pressures and user backlash, rather than a genuine commitment to ethical AI development. The consensus among critics is that it’s too little, too late, and that regaining public trust will be an uphill battle, if not an impossible one.