Today, the White House issued an executive order on safe, secure, and trustworthy artificial intelligence. We won't provide a full synopsis of the Executive Order here, but you can read the order in its entirety here.
We at Gleen applaud the Executive Order. Generative AI is powerful, and bad actors will use generative AI to do bad things.
That said, nothing in the Executive Order surprises us. In fact, Gleen AI already incorporates many of the key points of the Executive Order. Specifically:
On AI Safety
The Executive Order directs the Department of Homeland Security to set rigorous standards for extensive red-team testing.
Vendors must inform the government if their AI models threaten national security, economic security, public health, or safety. They also must share the results of all red-team tests.
All AI systems must protect against engineering dangerous biological materials.
- Will not threaten national security, economic security, public health, or safety.
- Cannot create dangerous biological materials.
On AI Security
Gleen keeps our customer's data, our infrastructure, and our software development processes secure. To this end, Gleen AI is in the process of becoming SOC 2 Type 2 certified.
On AI Trustworthiness
Trust is an integral part of Gleen's mission. In order for generative AI to be trustworthy, it can't hallucinate.
Gleen has focused (almost exclusively) on solving hallucination. As a result, Gleen AI is one of the most trustworthy AI solutions on the market.
The Executive Order also seeks to protect Americans from AI-generated fraud and deception. It hopes to establish standards and best-practices for detecting AI-generated content and authenticating official content.
Gleen AI currently focuses on text-based content, making it difficult to include watermarks within the content.
Gleen will continue to monitor technical solutions for brands that would like to mark its chatbot content as "official."
On AI Privacy
Gleen preserves the privacy of our customers' customers. Gleen AI automatically detects PII in any end user prompt. It then replaces/obfuscates the PII before sending the prompt to an LLM for embedding.
On Responsible AI
Gleen strongly believes in responsible AI solutions. We believe that the best customer support/success solutions marry AI and human efforts. Read more about Gleen's Generative AI Ethics pledge here.