‘Tech for good’: How a messaging app is improving public health in China

Everyone loves nan thought of personalised healthcare, but erstwhile it’s been trialled successful nan past, nan costs is often excessively high, some for nan pharmaceutical companies producing nan narcotics and for nan patients and healthcare systems. 

“That's why nan pharma companies for illustration to create ‘blockbusters’. It's conscionable 1 supplier for all,” Dr Alex Ng told Euronews Health. 

But what if AI could alteration each that?

We sat down pinch Dr Alex Ng, president of Tencent Healthcare, to perceive astir nan game-changing possibilities of AI successful healthcare and surviving up to nan company’s motto of ‘tech for good’. 

Access to healthcare

Whilst we each cognize exertion tin bring immoderate awesome innovations and improvement to an industry, Ng is very conscious of nan problems a deficiency of entree tin create arsenic a work becomes much digitised.

“What we person seen pinch history is that pinch digitisation, pinch much and much machines, you are really inadvertently widening nan inequality gap,” Ng explained. 

“A batch of that whitethorn beryllium owed to access, and a batch will beryllium owed to reimbursement, payments, [and] expertise to pay. Now, we've seen that pinch each azygous activity of exertion development, it gets wider and wider”.

Tencent is simply a Chinese exertion conglomerate but it is champion known for its celebrated messaging app, WeChat. 

By leveraging nan fame of nan app, Tencent person added other features to thief summation entree to healthcare for their users successful China. Users tin book infirmary aliases session appointments, get a tele-consultation and browse wellness and supplier accusation done their balanced of WhatsApp. 

“We utilized to constitute a batch of aesculapian articles to do wellness education, wellness awareness. But now, group are going to AI, whether it's really ChatGPT aliases chatbots,” Ng said.

“We tin make judge that nan reply that they get is overmuch much rigid and verified and overmuch much peer-reviewed, versus a batch of nan hallucinations that AI tin sometimes get”.

This is simply a peculiarly important work successful China, where, Ng explained, patients person their trial results astatine nan aforesaid clip arsenic a doctor. And for illustration astir of america would do too, galore patients look to nan net to construe nan results earlier their follow-up assignment pinch their doctor.

“A batch of nan time, we supply devices against AI to really explicate to them what these trial results mean. Of course, pinch nan accustomed caveat, that we cannot switch nan doctor's advice,” he explained. 

“Even if we didn't supply a rigorous AI motor to thief them pinch that, they would conscionable do it connected immoderate different random AI that is not healthcare specific. If they conscionable spell to a random instrumentality and they do it, nan mirage inaccuracy mightiness beryllium moreover higher. 

“And truthful we travel from a vantage constituent wherever we are good alert of nan seriousness of supporting nan diligent and nan personification successful interpreting immoderate of nan results. So we effort to connection a overmuch much crushed truth alternative”.

Can we judge AI that makes mistakes successful healthcare?

The results of utilizing AI for nan greater bully of nationalist wellness look evident but wherever nan statement gets absorbing is our capacity to judge mistakes. When moving pinch humans, we look to beryllium overmuch much accepting of quality error, yet expect AI, a human-created technology, not to make immoderate mistakes.

“AI is ne'er perfect. Just for illustration drugs, each supplier is intended for a definite indication, but you besides judge that it will person broadside effects and immoderate mild, immoderate moderate, immoderate serious, for nan full population. And there's a regulator, location are laws astir to protect that, because different nary 1 will beryllium investing successful developing caller drugs,” Ng said. 

“I deliberation nan aforesaid for AI. If we person 1 AI being developed for a circumstantial task, it mightiness beryllium amended than humans already, but it is not faultless. If location are definite faults wrong a boundary, that is acceptable, really do you modulate that? 

“How do we, arsenic a wellness system, activity pinch nan regulators, activity pinch wellness systems successful immoderate way, activity pinch nine connected what is acceptable? And I deliberation that statement is very different for China, Southeast Asia, nan US, Europe, and nan UK because I deliberation nan anticipation is very different”.