2. ChatGPT, other large language models will make certain attacks harder to identify.
Generative artificial intelligence is the capability of algorithms to automatically generate content from user queries such as text, video and images. It’s also a potential threat to data security, according to the Spiderlabs report.
While many companies have already begun investing in their own models, there were initial fears large language models would begin writing malicious code. That threat hasn’t yet fully materialized because the models require someone ro first understand how to code before the models produce malicious results, Sigler said.
But Sigler said AI is making it harder for employees to identify phishing scams or malicious emails.
“When you have an AI engine that speaks that language natively, and understands exactly what you’re trying to say, that becomes a lot more compelling,” Sigler said. “Those red flags, those grammatical errors, those spelling errors tend to disappear.”
The report also said healthcare systems may face an increased risk of exposure due to their reliance on third-party vendors that may incorporate generative AI into their products.
3. Healthcare lacks accurate inventory of devices.
The increased number of connected devices in healthcare further amplifies the vulnerability of the industry’s infrastructure, the report’s authors said. Devices ranging from an employee’s cell phone to medical equipment are all at risk.
Sigler said many providers don’t have an accurate inventory of connected devices.
“I see that as the biggest problem,” Sigler said. “Having a current, proper and ongoingly updated inventory of what you have, and how valuable they are to your organization…is going to help you prioritize the security controls you’ve put in place.”
Once an organization has an accurate inventory, Sigler said organizations should prioritize issuing a value to each area of data. Clinical data, for example, would likely have more controls and greater value within an organization than website analytics or marketing email lists.
“I think a lot of people just put the cart in front of the horse,” Sigler said. “They start setting up all kinds of policies and procedures…without actually understanding the complexities of implementing those policies.”
4. Personally identifiable medical information is available on the dark web
Stolen information potentially taken from U.S.-based healthcare organizations is likely available on the dark web, according to the Spiderlabs’ report.
There were 8,000 logs claiming to have information from U.S.-based healthcare organizations available on the RussianMarket forum, a popular underground marketplace.
For example, on CruptBB, a background forum where hackers sell information, there could be a healthcare attack advertisement that outlines the sharing of personal healthcare data such as medical records, social security numbers, phone numbers, addresses and names. According to Sigler, this is pretty typical of how advertisements on the underground forums look and the types of data offered for sale.