As artificial intelligence (AI) and the digital landscape grow, human interaction seems to be pushed to the wayside, but new programs encourage the connection between the two worlds. Mixus AI, a new AI platform used in Stanford classes, encourages discussions that involve both generative AI and human collaboration.
Elliot Katz, co-founder of Mixus AI along with Shai Magzimof, beta-launched the platform at the end of 2024 within Stanford, which is now being used in Lean Launchpad, an entrepreneurial course run by the School of Engineering (SoE) at the University open to all students.
Katz and Magzimof both have experience working with the integration of AI and human interaction. Before Mixus AI, they created Phantom Auto, a vehicle teleoperation technology company that empowered humans to remotely operate autonomous vehicles.
“We’re techno-optimists,” Katz said. “We believe in AI, but we also believe that you have to have a human in the loop.”
When using Mixus AI, users can create chats that include people and an AI platform of their choice (OpenAI, Perplexity, Anthropic or Google), blending artificial and human ideas. Katz explained how the incorporation of human interaction not only helps avoid the concern of AI inaccuracy but also fosters human connection.
“I believe that humans, by our nature, are connected animals, right? We want to connect with one another,” he said. “If AI is happening, we should do AI in a way where it augments humans’ abilities, and it augments humans’ abilities to still connect.”
Jeff Epstein MBA ’79, former chief financial officer at Oracle and lecturer in the Management Science and Engineering department, helped introduce Mixus AI to Lean Launchpad as part of the teaching team in winter quarter.
“The Lean Launchpad class is all about experimentation… we love the idea of experimentation,” Epstein said. “When we hear of a tool like Mixus, we said, ‘Why don’t we just run an experiment and see if the tool is useful for our students?’”
Epstein was drawn to Mixus AI’s ability to create group chats with one another, experts, and generative AI serving as an additional voice to a conversation. “There isn’t any existing product that has all these features combined, and I’m pretty excited about it. We’ll see how useful it is,” he said.
Lean Launchpad students have also enjoyed using Mixus AI to advance their business ideas. For example, Amit Shimshi MBA ’25 and her team use the AI platform to connect with other students for interviews in the e-commerce shopping industry.
A defining feature of Mixus AI that has been beneficial for Shimshi is the platform’s ability to recommend other users relevant to the conversation. “This has been a really great way for us to actually get in touch with students to interview,” she said.
Even though conversations about AI and education are still developing, Shimshi noted how the course has taken a more open approach to generative AI in the classroom.
“[The course is] absolutely encouraging for us to continue to figure out how to best use all these different AI tools to accelerate our learning,” Shimshi said. “It seems like professors really understand that this is the future, and we need to understand how to use these tools to boost ourselves.”
Epstein echoed this sentiment, explaining how Stanford gives faculty discretion to decide the best AI practices within their classrooms. With this, the Lean Launchpad team has decided to view AI as a learning tool for students. “AI in general is just the next version of Excel and Word and PowerPoint and all the other tools that we use,” he said.