When it comes to identifying and containing data breaches, organizations armed with security AI plus automation spend 108 days fewer than their differently prepared counterparts. The AI-wielding orgs also incur $1.76 million less in breach-recovery costs.

The findings are from an IBM Security analysis of data compiled by Ponemon Institute. Ponemon gathered the data by interviewing qualified representatives of 533 organizations that suffered a breach between March 2022 and March 2023.

Of note, IBM Security regularly speaks of “security AI and automation” as a single, unified thing. Going by language in a new report, the company defines the term as technology that “enables organizations to predict threats, protect data as it moves, and respond with speed and precision without holding back business innovation.”

Other key findings from the report, “Cost of a Data Breach 2023,” which was released the last week of July and is the sixth in the series:

• For the 13th consecutive year, healthcare shows the most expensive data breaches of any sector, shelling out an average of $10.93 million per breach. Far behind are runners-up financial ($5.90 million), pharmaceuticals ($4.82 million), energy ($4.78 million) and industrial ($4.73 million). The average across all industries was $4.45 million.
   
• That’s up 8.3% over 2022 ($10.10 million)—and a 53.3% spike over 2020, when healthcare reported an average breach-recovery price of $7.13 million. IBM notes healthcare’s rising average breach cost since the start of the COVID pandemic, adding that healthcare is both heavily regulated and considered critical infrastructure by the federal government.
   
• Slightly more than half of interviewed organizations, 51%, are planning to spend more on security in hopes of avoiding another painful bite in the back pocket. Top areas named for additional investments include incident response planning and testing, employee training, and threat detection and response technologies.
   
• 82% of all 2022-’23 breaches involved data stored in the cloud—public, private or multiple environments. IBM says these cyber break-ins incur a higher-than-average cost of $4.75 million.
   
• 61% of organizations employ some level of security AI and automation. Only 28% of organizations extensively used security AI and automation tools in their cybersecurity processes, while 33% had limited use, IBM says, adding: “That leaves nearly 4 in 10 relying solely on manual inputs in their security operations.”

Worth noting is that healthcare orgs comprised the smallest slice of the 17-industry study cohort.

Read a report summary from IBM’s marketing operation or the full report (in slides).

Buzzworthy developments of the past few days.

• President Biden has been noodling around with ChatGPT. The personal experimentation helps explain his fixation on the technology. (He talks about it quite a lot, doesn’t he?) The Wall Street Journal looked into the matter and found the head of state/commander-in-chief “amazed by the speed with which the tool can generate material.” The newspaper also spoke with Biden’s deputy chief of staff, Bruce Reed, who says his top boss views AI as “an existential challenge. This is a technology that has the potential for phenomenal good or ill.” Article here (behind paywall).
   
• AI has helped a quadriplegic move and feel physical sensations. The man lost those abilities three years ago in a diving accident. Now microchips implanted in his brain use the technology to translate thoughts into electrical signals so he can enjoy some control over his arms, hands and fingers. The man, a former NYC wealth manager, held a family member’s hand as a joyful act of triumph over tragedy. Great story.
   
• One way to gauge investor interest in AI is to count how many times it’s mentioned in quarterly earnings calls. Someone is actually doing that. And they’re finding the mentions way up. In an Intel call last week, for example, attendees uttered “artificial intelligence” or “AI” 58 times—up from just 15 in April. An Alphabet call had even more mentions, Microsoft was about even with Intel, and other major players were sitting similarly pretty on the scoreboard. Reuters has it.
   
• Thanks to the Freedom of Information Act, or FOIA, federal agencies get swamped with requests for records. At least three are seeing if AI can help lighten their load. Those would be State, Justice and the CDC. The tryout has supporters and detractors, of course. Either way one leans, it’s easy to empathize with agency staff straining to keep up or catch up. As NBC News reports, the 120 agencies under FOIA’s thumb collectively received more than 928,000 requests for information last year alone—while facing a backlog of well more than 200,000.
   
• Duke Health has set out on a five-year journey of AI discovery with Microsoft. The admired academic medical institution in Durham, N.C., says the partnership will focus on generative AI and cloud computing to “redefine the healthcare landscape.” Major planned yields include a new lab with innovation in its very name. Announcement.
   
• Would you believe there’s an AI-based system that predicts COVID-19 hospital admissions using little more than samples of wastewater? Believe it. Aussie researchers are making it happen—and with wastewater data from 100 million Americans living across 159 U.S. counties. The project is underway at the University of Technology Sydney. It’s described in scientific terms in Nature Communications and in lay language in a university news item.
   
• If random opinion-holders have it right, predictive analytics will outdo generative AI at helping healthcare be all that it can be. So found online pollsters at Becker’s Hospital Review when they posted a flash survey on LinkedIn. The sole question: “Which technology holds the most promise for healthcare?” Some 1,027 interested people weighed in. Predictive analytics topped the four-choice field at 41%. Generative AI wasn’t all that close, clocking in at 29%. Wearables (16%) and digital health apps (15%) brought up the rear. Becker’s own coverage includes some expert takes on the results.  
   
• Surgery augmented-reality company Augmedics (Chicago) has submitted the winning bid at auction for the digital health assets of Surgalign. The latter filed for Chapter 11 bankruptcy in June. The former raised $82.5 million in Series D financing that same month. Augmedics says it will seal the deal upon approval from a bankruptcy court slated for Aug. 8. Announcement.
   
• One in four Americans would not visit a healthcare provider who does not embrace AI. Or so they told surveyors from the practice-automation vendor Tebra of Newport Beach, Calif. The project drew responses from more than 1,000 consumers and 500 healthcare professionals. Summary findings here.
   
• CareCloud (Somerset, N.J.) is working with Google Cloud on bringing generative AI to outpatient clinics and other ambulatory practices. CareCloud notes the U.S. is home to around 48,000 such sites and likely counting.