A technological startup founded by a group of members of the US retired military service has received a patent for the new technology which aims to provide resistance to explanation and hallucination to artificial intelligence capacities.
By receiving the patent, which was issued Tuesday by the American Patent and Brands BureauThe leaders behind Data² hope to help open the AI black box to inject more confidence in the adoption and use of technology, especially in mission sets with high issues such as defense, intelligence and national security.
“These are organizations which, in the end, have a low tolerance for failure, in particular in defense and in Intel space. Bad answers have real consequences,” CEO Jon Brewton told Defenssoop. “But what it means is that they can finally create and deploy the AI in which they can trust and use to make critical decisions.
Brewton, a former aviator, founded Data² with Chris Rohrbach, a retired commander of Navy Seal, and Eric Costantini, a sailor, as well as Jeff Dalgliesh, with whom he worked in oil and gas, another industry that was the key to creating the company. In addition, the organization’s advisory team features members like the former Federal IOC Suzette Kent and Nancy Morgan, a former data director for the American intelligence community. At the beginning of 2023, when the group launched what was going to become data, they took the idea in the Catalyst Accelerator program with the Air Force research laboratory and the spatial force, and introduced proof of concept associated with the Intel community.
This expertise in the field was the key to the business course, in particular to understand the shortcomings he tries to fill for the services that Brewton and his co-founders served.
“We are counting on our expertise in the matter and our history to orient ourselves and restrict our concentration on what we really pay attention to,” said Brewton, adding that they are targeting “high reliability industry applications” in sectors such as defense, intelligence and finance.
He continued: “The Future We Trying to Build is one where a military commander on the field can trust the ai-gehed battlefield assessment, so cybersecurity analysts can rely on a ai to respond to threats in near real time. We’re really trying to make ai trustworthy for governments They can use it in Those Mission-Critical Areas, and So that People and Machines Can Really Start To Make Better Decisions Together… (and) so that they can start generating value from the data that they already have in this technology and this ability.
Brewton has explained that patented capacity has been designed to be technological agnostic and can work with any IA model or ecosystem, because it uses “knowledge graphic” technology in addition to the existing data of an organization, “the land setting of the AI in a fact base” so that “each response generated by AI is anchored in a verifiable and a source of access to a level of verification and data”.
Despite major progress, today’s fundamental AI models still have major limits that cause hallucinations and are often unreliable, which is a concern for defense officials.
For this reason, said Brewton, Data² has adopted a data -oriented approach to “resolve the lack of confidence, transparency and explanation”.
“In the end, what we discovered very, very early in the process, which informed the patent that we developed is that a better data architecture, a better data structure, is really the key to unlocking how you can develop in the explanation, traceability and transparency-not better models,” he said.
Currently, Data²’s commercial approach is to associate with larger technology suppliers, such as Amazon, Microsoft, Dell and Nvidia, with large footprints through which it can offer its services in the widest possible way with “brand equity” and an association with “some of the most reliable technological partners in industry, high -level technology suppliers to the American government.”
“It is not limited to data architecture, it is not limited to a data type, it is not limited to a wide language model, it is not limited to an environment,” said Brewton. “We have tested fully accommodated environments in the cloud, up to a completely open -air edge kit.”
Although the plan in the future is to continue to rely on this partner ecosystem to progress with the defense and intelligence activities, Brewton said that the acquisition of the patent “really justifies what we fully believed in a truly transformational approach in the way of developing AI, especially in these critical mission spaces”.
“Our patent really represents a kind of first step towards the creation of AI usable by governments and other highly regulated organizations,” he added.