“It’s a huge undertaking to try to create agreements to get organizations to give their data to those who are trying to create AI. That’s why AI in healthcare is moving much, much more slowly than it should be and needs to be. The press releases that you read would have you believe that it’s taking off, but it’s not. It’s moving slowly,” BeeKeeperAI CEO Michael Blum declared in a recent interview.
The press releases to which he’s referring, issued by companies like AWS, Google Cloud, Hippocratic AI and others, are certainly plentiful. But Blum said they misrepresent the actual pace of healthcare AI development — making people think that healthcare AI is having moment when, in fact, significant barriers still exist.
With the Rise of AI, What IP Disputes in Healthcare Are Likely to Emerge?
Munck Wilson Mandala Partner Greg Howison shared his perspective on some of the legal ramifications around AI, IP, connected devices and the data they generate, in response to emailed questions.
Some investors seem to agree with Blum and think his startup could help, as evidenced by the funding round his company announced on Tuesday.
The startup, which is focused on facilitating the development and deployment of healthcare AI models, raked in $12.1 million in Series A funding led by Santé Ventures. Other investors included the Icahn School of Medicine at Mount Sinai, University of California San Francisco (UCSF), AIX Ventures, Continuum Health Ventures and TA Group Holdings.
BeeKeeperAI was spun out of UCSF Health last year. Before the creation of the startup, Blum had been working at the university health system for about 20 years — he even created and led the UCSF’s Center for Digital Health Innovation. As part of that work, he and his team were building AI models to be used in healthcare settings.
Through these efforts, Blum realized that one of the most significant barriers to creating impactful patient-facing AI was developers’ lack of access to patient data. Healthcare delivery organizations are hesitant to share this data out of fear that they could be violating HIPAA or infringing on patients’ privacy, he pointed out.
A Personalized Approach to Medication Nonadherence
At the Abarca Forward conference earlier this year, George Van Antwerp, managing director at Deloitte, discussed how social determinants of health and a personalized member experience can improve medication adherence and health outcomes.
This issue is why BeeKeeperAI was founded. The startup offers a secure platform to help provider organizations feel more comfortable sharing their data with AI developers, as well as help those who are in need of data to build healthcare AI tools.
The company’s name represents a protected “hive” where the worker bees can create healthcare AI “honey,” Blum said.
“Our platform has created a highly secure computing space within Microsoft Azure. And in that computing space, data holders and algorithm developers can come together, and they can create, validate and deploy healthcare AI without putting patient data at any risk or the algorithm intellectual property at any risk,” he explained.
To protect sensitive data, the platform integrates with Azure’s confidential computing capabilities. Confidential computing refers to a cybersecurity concept that focuses on protecting sensitive data while it is being processed by computer systems — it aims to provide a secure and trusted environment for processing data, even when the underlying infrastructure or software may be compromised.
Traditionally, data is protected when it is being stored and when it is in transit, but when it is processed by applications or services, it becomes vulnerable to potential attacks or unauthorized access. Confidential computing addresses this gap by safeguarding data during computation, ensuring its confidentiality and integrity.
BeeKeeperAI’s also platform operates on a “zero-trust” model, Blum pointed out. This refers to a security approach that assumes no trust between different entities within a network, regardless of their location. It is based on the principle of continuously verifying and validating every access request and transaction, instead of relying on traditional perimeter-based security measures alone.
Blum described his startup’s business model as “a new marketplace where algorithm developers can access data sources that they never would have been able to access before.” He also pointed out that provider organizations that are sitting on a wealth of patient data “can pursue their mission of discovery and patient care more safely than they could previously.”
Both data stewards and AI developers pay a subscription fee to be on the platform. Novartis is one of BeeKeeperAI’s customers on the developer side of things, and UCSF and Mount Sinai are two examples of data-holders that are on the platform, Blum said.
In his view, BeeKeeperAI is different from other companies that offer healthcare data platforms because most other platforms require healthcare organizations to send them data. Blum claimed his startup is the only platform that allows the organizations to keep their data in their own protected cloud environment.
“With BeeKeeperAI, that data stays with the data holders and is always protected. It never leaves their control. The platform allows a developer to bring their algorithm to the data to either validate, train or deploy the algorithm without putting any data or intellectual property at risk. It protects both sides of the transaction,” he said.
Photo: metamorworks, Getty Images
