AI applications are growing in all industries and sectors. But how do we make sure that they can be deployed and used in a way that everyone can trust? After proposing its 7 key requirements, the High-Level Expert Group on AI developed the ALTAI list to facilitate business and organisations in self-assessing their AI systems. In this section, members of the AI Alliance share their experience with ALTAI or other practices that help in building an AI ecosystem of trust in Europe and around the world. 

Can't see the "add a post" button? Log in here and if you are on mobile scroll until the end of the page.

Man and woman scientist discussing in front of a supercomputer
© istock

EIT Health: AI and Ethics in the Health Innovation Community

An extensive survey was run by the EIT Health amongst startups and innovators. The survey is one-to-one with the AI & Ethics guidelines. In addition a set of case studies has been collected. All was presented and discussed at the World Health Summit in Berlin, 27 October 2019. An extensive report about the survey and cases and with recommendations is available. An op-ed was published in Science|Business.

1129803_eit_health_whs_report_final.pdf
(10.09 Mo - PDF)
Télécharger
Étiquettes
survey health ageing recommendations EIT health cases

Comments

Profile picture for user njastrno
Soumis par Norbert JASTROCH le sam, 11/23/2019 - 19:46

 Dear Paul Timmers,

The central problem, that of the balance between regulation and innovation, seems to be somehow similar to a problem discussed in A. Nassehi et al, The Strength of Weak Procedures, Zeitschrift fuer Soziologie 48(3), de Gruyter Oldenbourg 2019, https://doi.org/10.1515/zfsoz-2019-0015on the subject of organ donation. The problem they discuss in the context of organ allocation for transplantation is that of fundamentally diverging perspectives which make a specific question unresolvable in a general sense.

They argue that instead of trying to force obligatory political decision, it is more promising to organize an institutionalized permanent ethical reflection on the subject under consideration. 

Taking such an approach and transferring it to the issue of balancing regulation and innovation in the field of AI may offer a way how to deal with the fundamental perspective differences you address.

Regards, Norbert Jastroch

 

 

En réponse à par Norbert JASTROCH

Profile picture for user ntimmpul
Soumis par Paul TIMMERS le dim, 11/24/2019 - 09:24

Dear Norbert, very valuable comment, thanks! As regards the governance perspective and coming out of our survey and case studies is the practice of post-market surveillance in the medical devices and pharma field. Combined with pre-market impact assessment this suggests at least two governance mechanisms that can be institutionalized, i.e. mandated by law. These would well go together with the institutionalized ethical reflection that you refer to (checks & balances) in order to create more space for responsible and fast innovation.

Best regards, Paul

En réponse à par Paul TIMMERS

Profile picture for user njastrno
Soumis par Norbert JASTROCH le mar, 11/26/2019 - 18:32

Dear Paul,

one could expect these three elements - proactive regulation, reactive evaluation, and steady ethical reflection exercise - to build a promising innovation pathway. It's all about combining them economically reasonable, and institutionalize them in a prudent way. 

Kindly, Norbert 

 

Profile picture for user ngaursac
Soumis par Sachin GAUR le lun, 11/25/2019 - 02:42

Dear Paul, 

Thanks for sharing. I believe for healthcare especially regulation is needed to bring confidence among clinicians for uptake of AI. We have seen some very interesting examples of AI systems being used in parallel with the human experts. It is interesting to see the false positives and negatives in a clinical setting and the possible implication of the same on the patient's life if the human expertise was not available and vice versa. (so both approaches have their shortcomings and together they seem to be an improved system). 

I can share more with you offline exact examples on these AI vis-a-vis Human expert in clinical setting. 

 Thanks 

-Sachin