Tech Trends

AI-Mazing Tech-Venture: National Archives Promotes Google Gemini AI On Employees

By Mark McDonnell

National Archives Pushes Google Gemini AI on Employees

The US National Archives and Records Administration (NARA) conducted a presentation and tech demo ‘AI-mazing Tech-venture’ for the employees in June. In this program, Google’s Gemini AI was presented as a tool archive, that can be used by the employees to ‘enhance productivity’. During a demo, the AI encountered a question regarding the assassination of John F Kennedy. 

NARA intends to launch a public-facing AI-powered chatbot called ‘Archie AI’. A NARA spokesperson has announced that ‘The National Archives has big plans for AI’. This one is going to be crucial to decide how we conduct our work, how we should scale our services for Americans who want to access our records anywhere at any time, and how we guarantee that we are ready to take charge of the records being created today and in the future. 

Employee chat logs have been submitted during the presentation, which confirms that National Archives employees are concerned regarding the idea that AI tools will be employed in archiving. This is a practice that is inherently concerned with accurately recording history. 

One of the workers who attended the presentation apprised that ‘I suspect they are going to present it to the workplace. I’m someone who works there and hates AI bullshit’. 

This presentation happened after the National Archives banned employees from using ChatGPT, because it posed an ‘unacceptable risk to NARA data security’, and alerted employees that they should not depend on LLMs for authentic information. 

One of the slides noted that ‘Google Gemini is a versatile tool that can help users save time, improve productivity, and achieve better results in work’. Consider Gemini as a colleague who can help you to generate ideas and review your content. 

Some portions of the slides showing specific examples have been removed, however, the presentation suggests utilizing it for things such as writing assistance, creating data visuals, meeting summaries, and bringing up new ideas. 

One of the slides says ‘Generate text, translate languages, and summarize documents to help users communicate more effectively’. 

The presentation was conducted over a Zoom meeting and employees raised their concerns regarding the technology in the chat. The National Archives withheld the release of the chat, mentioning privacy concerns, the details regarding whether the presenter addressed the concern or how they answered remains a mystery. National Archives have redacted the names of the employees who have raised concerns/ questions. 

One of the questions raised by an employee is ‘How can someone be sure whether they are receiving a response from a real archivist or it is from a generative AI?’. The other two employees also stated that they are also concerned about the same. One of them has asked, ‘Would NARA disclose what aspects of reference are generated from AI?’, and they are also worried that, this might affect the trust in the institution if not properly disclosed. 

As per the chat logs, a live demonstration of Goole’s Vertex AI was given, it pretended to be an ‘expert archivist’ and was asked a question about the John F Kennedy assassination. Vertex AI allows organizations to train LLMs on their datasets. In the given case, the AI was trained on National Archives data. These queries included ‘Who killed Kennedy?’ and ‘What was the CIA’s involvement in the assassination of Kennedy?’. 

One employee has asked ‘Why is the Generative AI calling itself an ‘expert archivist?’. Someone who has been part of the demo replied that ‘It is called ‘expert archivist’ because that is the prompt we gave it’. 

Three employees have expressed their concern about the environmental aspects and carbon footprint of generative AI. 

National Archives told employees that they recommend Gemini and Microsoft Copilot to ChatGPT since they provide a more controlled environment. 

A NARA spokesperson has stated that the agency has a massive project for AI, that includes introducing a public-facing AI-powered chatbot, ‘Archie’. However, they are also planning to inform users that Archie may also provide incorrect information. 

The Biden administration has previously informed federal agencies to learn AI and create policies for its use. Recently, the National Archives presented AI to the International Council of Archives. 

NARA is developing ArchieAI, an AI assistant which is powered by Google Vertex. This will provide summaries with links to digitized documents. They are hoping to launch it soon. 

Further Reading: Google Turns to Small Nuclear Power Plants to Fuel AI Growth

Mark McDonnell

Mark McDonnell is a seasoned technology writer with over 10 years of experience covering a wide range of tech topics, including tech trends, network security, cloud computing, CRM systems, and more. With a strong background in IT and a passion for staying ahead of industry developments, Mark delivers in-depth, well-researched articles that provide valuable insights for businesses and tech enthusiasts alike. His work has been featured in leading tech publications, and he continuously works to stay at the forefront of innovation, ensuring readers receive the most accurate and actionable information. Mark holds a degree in Computer Science and multiple certifications in cybersecurity and cloud infrastructure, and he is committed to producing content that reflects the highest standards of expertise and trustworthiness.

Leave a Comment