AI Ethics Under Fire: OpenAI's Deal with the US Military Raises Concerns
The AI landscape is shaken by a controversial partnership. Just 7 minutes ago, Chris Vallanceand and Laura Cress, technology reporters for AFP, broke the news: OpenAI is reevaluating its agreement with the US military after facing intense backlash.
The original deal, described as 'opportunistic and sloppy' by OpenAI itself, sparked debates about the ethical use of AI in warfare and the balance of power between governments and private companies. OpenAI's statement on Saturday revealed a revised agreement with the Pentagon, boasting more safeguards than any prior classified AI deployment, including Anthropic's.
But the story doesn't end there. On Monday, OpenAI's CEO, Altman, took to X to announce further adjustments. These changes ensure their system won't be intentionally used for domestic surveillance of US citizens and require intelligence agencies like the NSA to modify their contracts before accessing OpenAI's technology.
Altman admitted the company's haste in releasing the initial agreement, acknowledging the complexity of the issues and the need for clear communication. He stated, "We wanted to de-escalate, but it seemed opportunistic and hasty."
The backlash was swift, with users uninstalling ChatGPT in droves after learning of OpenAI's Pentagon collaboration. Sensor Tower data shows a 200% spike in uninstalls compared to regular rates. Meanwhile, Anthropic's Claude AI model, banned by the Trump administration for refusing to develop autonomous weapons, has been secretly used in the US-Israel war with Iran, as reported by CBS News.
AI's role in the military is multifaceted. It's used for logistics, data analysis, and rapid information processing. Palantir, an American company, provides AI-powered tools to the US, Ukraine, and NATO for intelligence gathering, surveillance, and military operations. The UK Ministry of Defence recently signed a £240m deal with Palantir, integrating its AI platform Maven into NATO's systems.
However, AI models can err or even fabricate data, a phenomenon known as 'hallucinating'. Lieutenant Colonel Amanda Gustave, NATO's Task Force Maven chief data officer, assured human oversight, stating that AI would never make decisions without human intervention.
While Palantir advocates for human involvement in AI weapon systems, Anthropic pushes for a complete ban on autonomous weapons. With Anthropic's absence from Pentagon deals, Oxford University's Professor Mariarosaria Taddeo warns of a potential safety gap, stating, "The most safety-conscious player is now absent."
This BBC AI Unpacked week, we delve into the complex world of AI, its potential, and the ethical dilemmas it presents. Stay tuned for more insights and join the conversation!
