- Industry data will be crucial for the next stage of AI evolution
- Expect data alliances to form, both between model developers and partners as well as between specific industry groups
- Integration across datasets will also be key, OpenText told us.
You’ve heard of artificial intelligence (AI) model libraries. But what about data pools? As large language models begin to diversify into smaller, specialized iterations, access to industry-specific data will be critical to making them relevant, OpenText CMO Sandy Ono said to Fierce.
“That’s what’s valuable to the customer,” Ono said of the ability to integrate AI models into specialized data sets. And she knows this, given that OpenText is in the information management business and therefore well-versed in what enterprise customers want.
Consider a healthcare client. Their AI model “doesn’t need to mine the entire internet, but it mines all the healthcare laws and regulations in my country. So when I make decisions or generate the next step, it takes that into account.”
So where will these data sets come from? In a word, alliances.
Partnerships on the data front are already starting to emerge. Indeed, we reported earlier this week on Extreme Network’s initiative to power data from Intel devices in its AI network.
And late last year, the dominant force OpenAI launched a data partnerships program, through which it works to create open-source and private datasets for training AI models. On the private side, these partnerships give it to access to non-public information such as archives and metadata.
“We are interested in large-scale datasets that reflect human society and are not already readily available online to the public today,” OpenAI explains. wrote at the time.
Google appearis set to work on a similar data partnership initiative to advance its AI.
But Ono said vertical players could also form their own alliances.
“In any industry, you inherently know where to go,” she said. “In insurance, you have GuideWire and other companies like that. So data providers are going to start looking at these issues and forming alliances where it matters.”
There is one more piece to the puzzle: integration.
“Five years ago, you integrated applications because you wanted the workflow to be fluid. Today, we need the data to be fluid so that AI can work on it,” Ono said. “Customers will be the drivers, because they’ll be the ones saying, ‘Well, why isn’t this connected to that? Why do I have to go to another library?’”
“That’s one of the things I think we’re going to be moving toward over the next three to six months,” she concluded.
