Who Will Control the Future of AI?

AI is increasingly embedded in everyday life, but be mindful of how you use it

Artificial intelligence (AI) continues to revolutionize various sectors, with recent developments highlighting both its transformative potential and the emerging challenges it presents.

In this space, disruption happens quickly, so we all must pay attention.

Last week, DeepSeek launched its new AI model, creating shockwaves across the AI landscape because it is open source (freely available) and reportedly has a dramatically lower training cost. Concurrently, the health care sector faces increasing cybersecurity threats, particularly from sophisticated AI-driven attacks such as deepfakes.

Here are a few things to know as we start what is likely to be another eventful year in AI across sectors.

DeepSeek's Rapid Emergence

DeepSeek has garnered significant attention with the launch of its AI model, DeepSeek-R1. This model has quickly risen to prominence, surpassing established platforms like ChatGPT, to become the top application on Apple's U.S. App Store.

Developed with less investment and utilizing lower computing power, DeepSeek-R1 challenges existing AI paradigms and has prompted concerns among U.S. tech giants regarding their competitive standing.

The model's efficiency and cost-effectiveness has led businesses to explore its potential applications. Companies such as New York Life and SAP are testing DeepSeek's technology, attracted by its lower training costs compared to U.S. counterparts.

Importantly, data privacy concerns, especially related to Chinese ownership, have made some organizations cautious. If you recall, these are the same concerns that we previously had with TikTok.

Cybersecurity Threats in Health Care

As AI technology advances, health care faces heightened cybersecurity risks. The U.S. Food and Drug Administration (FDA) recently identified cybersecurity vulnerabilities in certain patient monitors manufactured by Contec and Epsimed. These vulnerabilities could allow unauthorized access, potentially leading to device malfunctions and unauthorized export of sensitive patient information. While no incidents have been reported to date, the FDA has urged health care facilities to implement mitigation strategies.

Deepfakes, AI-generated synthetic media, pose another significant threat to health care security. Cybercriminals can use deepfakes to impersonate hospital executives or other personnel, facilitating fraudulent activities and unauthorized access to sensitive information. The sophistication of deepfake technology makes it challenging to detect, increasing the risk of successful cyberattacks.

Mitigation Strategies

To address these challenges, health care organizations must adopt comprehensive cybersecurity measures. This includes regular software updates, employee training on recognizing phishing attempts and deepfakes, and implementing robust access controls. Staying informed about emerging AI technologies and associated risks is crucial for maintaining security and protecting patient data.

If your health care facility does not have an acceptable use policy, it is highly recommended to initiate the development of such a policy. This policy will identify what can be used internally and what should not be used by staff. It is also highly recommended that you never input sensitive information into any AI chatbot as this information could be stored in the AI algorithms and accessed or utilized by others.

Always read the terms of use when it comes to utilizing AI services; this extends beyond chatbots, but also to popular note-taking services.

Most of these AI note-taking services state in their terms of use that any meeting the AI is active on stores all of the discussion and essentially becomes the property of the AI company.

So, if you do utilize these AI note-taking apps, use caution, especially for meetings with sensitive content.

While AI developments like DeepSeek offer promising advancements, they also underscore the need for heightened awareness and proactive measures to mitigate associated risks, particularly in critical sectors like health care. For additional information about cybersecurity and health care, do not hesitate to contact me.

 

Author: CJ Sabo, emergency preparedness manager, The Hospital and Healthsystem Association of Pennsylvania

Return to Blog main page