Healthcare businesses urged to address these five AI risks
Amongst the risks are bias, misinformation, and privacy.
Whilst AI adoption continues to grow, BCG cautioned healthcare businesses against risks in using the technology.
In a study, BCG said providers and other health care organizations must ensure that generative AI is implemented safely and they could address the following:
The first is bias because outputs generated by AI can reflect inherent biases in underlying data and companies using AI systems need to assign experts to review data and results. This would mean correcting any imbalances.
Another is misinformation as AI models are still improving, which means sometimes they create results and other incorrect statements.
“To build trust, providers will need to ensure that outputs are subject to human review,” said BCG.
The third is privacy as firms using generative AI tools should clarify data ownerships with stakeholders, strengthen cybersecurity, and develop synthetic data.
The fourth is addressing the black box nature of generative AI systems. Healthcare firms can build trust by being transparent on how their algorithms work and how a specific data leads to a prognosis.
The last is misuse of generative AI outputs that should not be taken as final word.
“Hospitals, clinicians, and payers should make it clear to patients that AI-generated insights are recommendations, not definitive information,” said BCG.