- AI has a huge promise to
improve the quality in healthcare
and transform the way
we take care of patients
in a similar way that has
transformed many other industries.
Healthcare is different though,
because the depth and the variety of data
that you need to develop AI models
that can be really impactful
while being very safe
is very large.
AI developers need access
to those very large,
very deep data sets, yet the
healthcare organizations have
to protect that data.
Patients' clinical records
are incredibly sensitive,
and we very rarely share them.
Unfortunately, that's exactly
the data that you need
to share with an algorithm developer.
Microsoft with confidential
computing in Azure, and Intel
with their software guard extensions
were all coming together at
exactly the right time for us.
So it allowed us to develop a platform
where data stewards can keep
the data in their existing
protected HIPAA compliant clouds.
- At BeeKeeperAI, we're really
excited about the impact
of the platform because it accelerates
the development of
healthcare AI tremendously,
so that we can get algorithms
from the research lab
to the bedside quickly.
The reason that happens
is because the security
of the BeeKeeper platform
allows us to get access to data,
and run validation very, very quickly,
cutting the time down from two
years down to several months.
When we think about what
that means for healthcare,
we're talking about new
ways of diagnosing patients,
new ways of monitoring treatment,
of coordinating care for patients.
It's actually using the
Intel SGX technology,
which allows us to run
calculations inside almost like
a vault inside the memory of the computer,
so that once the data and the
algorithm are together inside
that secure enclave nobody
else outside can see
what's happening.
The algorithm can run on the data
in the clear inside that vault,
and then a validation
report can be created.
The unique thing about BeeKeeperAI
is that it allows algorithm developers
to validate their algorithms on data
that they don't directly have access
to without ever sharing the data,
and without compromising
the IP of the algorithm.
- It's always been about
improving the delivery
of care to the patients.
AI can really help us deliver better care,
and improve workflows,
deliver the right care
at the right time to a patient,
and you can make life saving
decisions much more quickly
than you can right now.
(soft music)