Stability AI's CEO on private data and looking ahead

Forty percent of C-level executives say their org will increase how much they invest in artificial intelligence, according to McKinsey research.

• Why it matters: As AI evolves, leaders need to not only be open-minded, they need to pre-plan strategies to help them navigate it. Otherwise, they risk falling behind competitors.

Emad Mostaque, Stability AI's CEO, told Axios HQ in a recent event that Stability AI has plans for "every model, every modality, every sector." They just released their Japanese language model trained from scratch and have plans to continue expanding into all areas — and other organizations should be ready to act on AI too. 

• “We see the wave coming. Now this time next year, every company has to implement it — not even have a strategy. Implement it," Emad said.  

And implementing it starts by knowing what’s happening now and what’s coming next.

Right now, private data is key

As AI advances and can do things like pass the bar exam, it’s not simply about the technology itself and how smart it can get. It’s about what it can do with the data companies presently have. 

“Creating models of every type and enabling the private data of the world is what's most important,” Emad said — because private data is going to be a key factor in how organizations use AI.  

Emad’s take on how Stability models could work:

• Current data usage: Companies have sensitive data they may not want to share in a cloud environment. That data often ends up unsearchable or unusable as a result. And yet, Emad considers private data to be “the most valuable to companies, and they want to turn that into intelligence.” So companies need to find a safe, actionable way to capture all that their data can offer.

• Add Stability models: Stability services examine all your documents, meeting notes, etc. to create a personalized model. Like a personal assistant, it can bring you knowledge from the people you’ve talked to, the emails you’ve received, or the client information you’ve gathered. Use that to save time, draw new conclusions, and advance your capabilities. 

• Expand your data: In the future, organizations will implement hybrid models for less private data, where organizations partner with ones like Stability AI — and they provide the models that the individual partners use to compliment AI services that they own to produce individualized results. 

“You can use our models without working with us, and it will transform your data into your own version of a chat model,” Emad said. But that’s what they consider their role in the ecosystem — open models to private data. Now it’s just up to organizations to make the leap and start using AI to their advantage. 

Companies need to start looking ahead now

Right now, it may all be about data, but Emad likens AI progress to being at the “iPhone 2G to 3G point,” meaning advancements have been made — but more is coming. Fast. 

A look at what Emad thinks could be coming for AI:

• Education: "In 10 years, every single student in the world will have an AI tutor that helps them," Emad said. Teachers will need to have AI plans in place — whether that's embracing it and having students create study guides with it or choosing to reject it. 
• Government: "Every government in the world will deploy this," Emad said, and it will have to be done in a way that is open and auditable. 

• Coding: “I think the one area we’ve seen the most impact right now is coding. In studies by Microsoft, it shows a 50% improvement in coding times. That’s almost doubling — because every coder should use a copilot or a stable code type model because it makes coding better.”
• Private companies: Data — using it and protecting it — is key. But the way leaders build leverage is by finding a way to use AI that turns their private data into usable intelligence. 

The bottom line: Organizations need teams in place who can make the most of the current state of the technology if they really want to position themselves for future success. “Just have a team that uses it, has an open mind, and thinks about where can I implement this in my processes to reduce friction and take it forward.” It’s the companies that start experimenting and advancing now that will likely come out on top. 

Read next: Generative AI is your assistant, not your replacement