Health IT

The downside of healthcare’s hottest buzzword: AI

A new Accenture survey notes that although healthcare execs are investing in AI and IoT, 81 percent of them said they aren’t prepared to face the societal and liability issues that may require them to explain their AI’s decisions and actions.

Artificial intelligence and the Internet of Things continue to come up in healthcare discussions and at prominent medical conferences and events. A new report from Accenture found that while executives are taking a keen interest in these technologies, there are some crucial issues to keep in mind before implementing them.

Accenture’s Digital Health Tech Vision 2018 report includes responses from 100 healthcare executives and directors at companies with annual revenues of at least $500 million.

According to the survey, 77 percent of respondents said they anticipate investing in IoT and smart sensors this year. Fifty-three percent expect to invest in AI systems.

Plus, 80 percent believe that within the next two years, artificial intelligence will work next to humans in their organization.

But it’s not all perfect. Eighty-one percent of surveyed executives said they aren’t prepared to face the societal and liability issues that may require them to explain their AI’s decisions and actions. That’s why 73 percent said they plan to develop internal ethical standards for AI to make sure their technology acts responsibly. That’s probably a good idea, given that 24 percent of executives said they have been the target of adversarial AI behaviors, including bot fraud or falsified location data.

With all the new tech, health systems also have to beware of a separate type of vulnerability: inaccurate, biased and manipulated data. This information can lead to distorted results. About 86 percent of respondents haven’t invested in tools to verify data sources from their various systems.

“As this paradigm-shifting technology [AI] evolves — making business more dynamic than ever before — organizations will remain responsible for demonstrating data stewardship and designing systems with trust and transparency to bolster the societal benefit of these technologies,” Kaveh Safavi, head of Accenture’s global health practice, said in a statement.

Despite these pitfalls, it seems AI will play a role in the medical world. A 2017 Accenture report, Artificial Intelligence: Healthcare’s New Nervous System, outlined which applications will have the biggest near-term effect on healthcare.

Accenture estimated each item will have the following value (or potential annual benefits) by 2026:

  • Robot-assisted surgery — $40 billion
  • Virtual nursing assistants — $20 billion
  • Administrative workflow assistance — $18 billion
  • Fraud detection — $17 billion
  • Dosage error reduction — $16 billion
  • Connected machines — $14 billion
  • Clinical trial participant identifier — $13 billion
  • Preliminary diagnosis — $5 billion
  • Automated image diagnosis — $3 billion
  • Cybersecurity — $2 billion

This adds up to a total value of nearly $150 billion.

As the report points out, AI for robot-assisted surgery can improve the physician’s instrument precision. Virtual nursing assistants can prevent nurses from unnecessarily having to visit patients and can keep patients from unnecessarily going to the hospital. And administrative workflow assistance applications can save clinical staff time by cutting down on non-patient care tasks like writing prescriptions.

Photo: MF3d, Getty Images