Walmart, Meta, and LinkedIn are among the companies that are currently exploring the implementation of internal generative AI solutions for their employees. These options prioritize the security and confidentiality of company data. The approaches taken by these companies vary, with some offering generative AI “playgrounds” that provide a range of models for employees to utilize, while others, like Meta, have developed their own proprietary internal chatbot.
These initiatives demonstrate a different stance compared to companies such as Goldman Sachs, Amazon, and Verizon, which have prohibited the use of publicly available generative AI tools like ChatGPT. The concern in these cases is likely rooted in the potential risks associated with utilizing external AI platforms that may compromise data security or privacy.
By opting for internal generative AI solutions, Walmart, Meta, and LinkedIn are striving to strike a balance between empowering their employees with AI capabilities and safeguarding sensitive company information. These companies recognize the importance of leveraging AI technologies to drive innovation and enhance productivity within their organizations, while also ensuring that data protection remains a top priority.
Walmart announces a new generative AI playground
Last week, Walmart announced its new Generative AI Playground, a platform the company describes as an “early-stage internal GenAI tool where associates can explore and learn about this new technology, while keeping our company and its data safe.”
The news builds on an interview in April with Desiree Gosby, VP of emerging technology at Walmart Global Tech, who told VentureBeat that the retailer is building on OpenAI’s GPT-4, among other models, and that generative AI is “as big a shift as mobile.”
The announcement, made last week in a LinkedIn post by Cheryl Ainoa, EVP of new businesses and emerging technology at Walmart Global Tech, says, “there will be various GenAI models available to try out all in one place…enabling our associates to see the difference in how each model reacts to the same prompts.”
The screen welcoming associates to Walmart’s Generative AI Playground says that employees “learn best from trial an error” and that the tool is a “safe way to try how GenAI can be used without risk of data leakage or exposure” and a place associates can use “more realistic prompts for their job function.”
It also emphasized that the tool is only for experimentation and work purposes; that its results have not been validated as accurate and “should be validated before being internally shared or used to help make business decisions.”
Meta employees get access to internal chatbot
The Verge reported this weekend that Meta has built an internal AI chatbot called Metamate that uses internal company data. The chatbot allows employees to create their own prompts and share them with colleagues. According to the report, Meta is “starting to roll it out internally to a small group now” to experiment in summarizing meetings, writing code and debugging features.
The Meta news comes on the heels of an internal all-hands meeting last week in which CEO Mark Zuckerberg said Meta is building generative AI into all of its products, including LLM-powered AI agents with unique personas and skill sets that help and entertain people. In addition, he highlighted an Agents Playground, an experimental internal-only interface powered by LLaMA where users “can have conversations with AI agents and provide feedback to help us improve our systems.”
Last month, Meta announced an AI “testing playground” for a small group of advertisers to use as they try out new generative AI-powered ad tools including text variation, background generation and image outcropping.
LinkedIn speeds up development with internal sandbox
In April, LinkedIn’s head of data and AI, Ya Xu, told VentureBeat about the company’s Generative AI Playground, an internal developer sandbox. The tool allows engineers to explore LinkedIn data with advanced generative AI models from OpenAI and other sources. The company also brought together engineers for LinkedIn’s largest-ever internal Hackathon, featuring thousands of participants.
Xu said that her team early on prioritized an engineering philosophy “rooted in exploration over building a mature final product.” The maturity for the right features and experiences would occur over time, she explained, but the exploration was encouraged by putting generative AI technology in the hands of every engineer and product manager who was interested.
“They need to have a better understanding of what the are problems are,” she said, “so when we build the product [using these models] we can prevent them from from happening.”
