Interestana
Home/News/SEC Officials Raised Alarm Over Amazon's AI Data Training Policies
Fast Company••4 min read

By Interestana AI Editorial — AI-drafted, human-overseen. How we report

SEC Officials Raised Alarm Over Amazon's AI Data Training Policies

SEC Officials Raised Alarm Over Amazon's AI Data Training Policies

In early 2024, officials from at least two federal agencies, including the Securities and Exchange Commission (SEC), experienced a period of significant anxiety regarding Amazon's updated terms of service. The core of their concern revolved around the potential for Amazon to utilize government data for training its artificial intelligence (AI) models. This apprehension was particularly acute for an IT official within the SEC's Cloud Center of Excellence. This individual, after meticulously reviewing Amazon's revised terms and accompanying industry articles, identified a critical clause that suggested the company might be permitted to use data residing outside a client's direct environment, and even beyond U.S. geographical borders. The official promptly issued an urgent warning, urging federal agencies to actively "opt out" of this data usage provision, labeling it an "urgent potential data leakage." The underlying fear was that Amazon could leverage sensitive government information to enhance its AI capabilities, a scenario with profound implications for national security and data privacy.

The scope of this potential issue was substantial, with the official identifying approximately 22 Amazon services that could be affected. Among the most prominent were Amazon Rekognition, a sophisticated image analysis service known for its facial recognition capabilities, which has previously been employed by the Federal Bureau of Investigation (FBI). Other services of concern included Amazon Polly, used for text-to-speech synthesis; Amazon Translate, for language translation; and Amazon Comprehend, a natural language processing service. The IT official's recommendation was clear: federal Chief Information Officers (CIOs), the highest-ranking technology decision-makers within government agencies, should engage directly with their Amazon account teams to advocate for an "official policy change" that would explicitly prohibit the use of government data for AI training.

This incident underscores a broader trend: the increasing reliance of federal agencies on major cloud service providers like Amazon, Google, and Microsoft. These tech giants are locked in a fierce competitive battle for lucrative federal cloud contracts. For context, Amazon had previously announced a substantial investment of up to $50 billion in government computing services, signaling its aggressive pursuit of this market. The SEC official's proactive stance and the subsequent inter-agency communication highlight the complex challenges federal entities face in navigating the intricate terms and conditions of these powerful technology companies. It raises critical questions about whether government officials always possess a complete understanding of the agreements they enter into, especially in rapidly evolving fields like cloud computing and AI, where data governance and privacy are paramount. The brief but intense concern serves as a stark reminder of the need for robust oversight and diligent review of service agreements to safeguard sensitive government information.

Original source — read the full reporting at the publisher:

Read on Fast Company

Get the weekly AI digest

AI news + new model releases, weekly. Drafted by our agents, reviewed by humans.

Read next