八点最热报 | AI人工智能远程医疗系统的背后，需要大量的数据作为支撑，只有通过丰富、多样的临床数据，AI人工智能系统才能不断学习、改进，从而实现更高效的医疗服务和个性化健康管理。相比起人力收集，利用AI人工智能远程医疗收集的数据真的会比较准确吗？
(主播：梁宝仪)
(主播：梁宝仪)
Category
🗞
NewsTranscript
00:00Before watching the video, let me remind you that there is more content on the Hotpoint website.
00:06In the field of medical health, the application of AI is gradually becoming popular.
00:11It also has a deep impact on the future of the industry.
00:14For example, AI is far-reaching in medical care.
00:16Through data analysis, model recognition, and automated decision-making,
00:20it has a strong ability to improve the accuracy of diagnosis.
00:24In terms of optimizing treatment plans and improving patient care quality,
00:28AI has shown great potential.
00:30But behind the AI system, a lot of data is needed as support.
00:35Only by enriching the clinical data,
00:38can the AI system continue to learn and improve,
00:42to achieve more efficient medical services and personalized health management.
00:46Compared to the data collected by manpower,
00:48will the data collected using AI remote medical care
00:52be more accurate in diagnosis?
00:56Professor Lv Ziqiang of the Department of Artificial Intelligence,
00:59said that the accuracy of the AI remote medical system
01:02mainly depends on two main factors.
01:05The first is the perfection of AI algorithms.
01:08Lv Ziqiang said that AI algorithms need to be constantly improved
01:12to further improve the accuracy of the model itself.
01:16Another factor is whether the data is sufficient.
01:20He said that in the absence of data,
01:23it is actually very difficult to train an AI model with high accuracy.
01:28The chief executive of UMCH, Zhang Shengren, added that
01:32a large amount of data means that
01:34the training of AI remote medical models
01:37requires sufficient patient support,
01:40especially the health data of Chinese patients.
01:43Zhang Shengren said that
01:44the AI models cultivated by the use of diverse health data
01:47can really be applied to Chinese patients.
01:53Basically, we need to collect some data
01:56and then train some basic models.
01:59So these basic models have not been unchanged.
02:02Because we know that there will be a lot of new data coming in.
02:06Depending on the situation of different patients,
02:09or the scene, it will be different.
02:12There are a lot of data out there.
02:14If we train it,
02:15it may not be suitable for us Malaysians.
02:18Because their eating habits,
02:19their living habits are different from ours.
02:22So if we train this model,
02:23it can't be used directly on Malaysians.
02:26The core of AI remote medical models
02:29is to use a large amount of data to train AI models.
02:32By analyzing and learning historical data,
02:35AI can better understand patients' health conditions,
02:38disease patterns, and treatment effects,
02:40thereby providing more personalized medical services.
02:43However, under the difference in the background of
02:46Malaysian multi-ethnic living environment, lifestyle, and eating habits,
02:50it is not suitable to train models using existing foreign data.
02:54Chairman of UMCH, Chairman Zhang Shengren, said
02:57that AI remote medical systems must be localized
03:01in order to truly meet the health needs of Malaysians.
03:05Of course, you have to be localized.
03:07You can't just use foreign data to implement.
03:10Many companies use foreign data to implement.
03:12The results may be different.
03:14We have to target different ethnicities.
03:16We have to target different regions.
03:18It's like you're in a city or a rural area.
03:22It's very different.
03:24We have to target different regions to collect the data,
03:27and then do experiments and reference.
03:30Professor Lin Liling of the Department of Medical Science, UMCH, pointed out
03:33that the user friendliness of AI remote medical platforms
03:36is also an important consideration in the design.
03:38The research team will work with patients' age,
03:41education level, and the ability to master electronic devices
03:44to further optimize and improve the value of the platform.
03:47When we develop these remote medical methods,
03:50we have to make it user-friendly.
03:53We also try to make it health literacy appropriate.
03:56You can't just say that remote medical is only suitable for those
03:59who receive higher education.
04:00In terms of user interface,
04:03it's also important to design a system that interacts with people.
04:06When designing, we have to consider different age groups
04:09and their knowledge of these IT devices.
04:12Secondly, do they have these related devices at home?
04:16For example, Wi-Fi.
04:18Is its screen able to support an AI platform?
04:22We have to target different communities and different age groups
04:26to continuously improve the design
04:28and enhance its practical value.
04:30Although the design has put in a lot of effort,
04:32the AI remote medical platform still needs the assistance of some devices
04:36to collect the patient's health data.
04:39In this way, the users of AI remote medical platforms all said
04:42that although the platform provides a lot of convenience,
04:45the price of testing equipment is not affordable for everyone.
04:49It's a big help.
04:51But this thing I have to buy,
04:52and it costs me about $250 to $300 every two weeks,
04:57which I can afford based on my pension.
04:59It needs to become cheaper because for me,
05:02$300 every two weeks is affordable,
05:05but not many people can afford that.
05:07I think we are quite good.
05:09Only perhaps if you can bring the cost down for the strips,
05:13or perhaps for the diabetes,
05:15you know, those new technologies available for the people
05:20who are not well and that would be appreciated.
05:25In addition, as technology advances day by day,
05:27new technologies keep emerging.
05:29Professor Lu Ziqiang of the Department of Artificial Intelligence,
05:31the University of Malaysia,
05:32believes that in an era where medical knowledge is constantly being updated,
05:35AI must have the ability to continue learning.
05:39In order to keep up with the current trend,
05:41Professor Zhang said that the team can only keep up with the times,
05:45and keep improving the long-term medical platform.
05:48When we design this AI model,
05:50we must ensure that AI can continue to learn,
05:54to adapt to new data,
05:58to adapt to new medical knowledge.
06:01So this model itself must have some flexibility
06:04to be able to keep up with and improve.
06:07At present, AI must have this kind of ability.
06:11As technology advances,
06:12these products are getting faster and faster,
06:14just like our mobile phones.
06:16You may have bought a new mobile phone,
06:19and after three and a half months,
06:20you have launched a new model,
06:22with different functions added to it.
06:24So we also need to keep updating our products
06:27and follow the hardware.
06:30AI in Malaysia
06:33Is the development of AI in Malaysia
06:35in a backward state?
06:38UMCH CEO Zhang Shengren said,
06:41UMCH and the University of Malaysia
06:43have been cooperating on AI remote medical care
06:46for 10 years.
06:48In this short period of time,
06:51they have successfully cultivated a lot of AI
06:54and medical expertise,
06:56and successfully launched WeHealth,
06:58a remote medical platform.
06:59Although the development of AI requires a lot of money
07:02and talent,
07:03technically,
07:04it is also a certain gap from other advanced countries.
07:07But Zhang Shengren emphasized,
07:08under the strong push of the government,
07:10the AI field in our country
07:12has entered the right track,
07:13and is moving in the right direction.
07:26For more UN videos visit www.un.org