Moti­on Mass­hardt (25.3210): Crea­ti­on of a natio­nal ethics com­mit­tee for arti­fi­ci­al intelligence

Sub­mit­ted text

The Fede­ral Coun­cil is ins­truc­ted to con­ve­ne an ethics com­mit­tee in the field of arti­fi­ci­al intel­li­gence (AI).

Justi­fi­ca­ti­on

AI systems are being used in more and more are­as – both in public and pri­va­te life. Within seconds, publicly acce­s­si­ble tools such as ChatGPT crea­te texts, images or pie­ces of music. AI appli­ca­ti­ons are used for lear­ning, faci­li­ta­te work pro­ce­s­ses and crea­te art. In addi­ti­on to the gre­at poten­ti­al ethi­cal que­sti­ons con­nec­ted: What does AI mean for poli­ti­cal opi­ni­on-forming & par­ti­ci­pa­ti­on when pro­pa­gan­da can be text­ed in unli­mi­t­ed quan­ti­ties with almost no effort? What effect do plat­form algo­rith­ms have on the relia­bi­li­ty of infor­ma­ti­on that peo­p­le con­su­me? What impact do AI appli­ca­ti­ons have on the world of work, art & cul­tu­re and edu­ca­ti­on? How can dis­cri­mi­na­ti­on be pre­ven­ted when AI is used in recruit­ment pro­ce­s­ses, for exam­p­le? How can we defend our­sel­ves against decis­i­ons made by AI systems? What pro­ce­s­ses are nee­ded to ensu­re respon­si­bi­li­ty and accoun­ta­bi­li­ty? To what ext­ent should AI appli­ca­ti­ons be used for sur­veil­lan­ce? And what rest­ric­tions on fun­da­men­tal and per­so­nal rights are we pre­pared to accept? Which peo­p­le & orga­nizati­ons crea­te, deter­mi­ne and con­trol AI systems?

Frame­work con­di­ti­ons are nee­ded to ensu­re that AI actual­ly bene­fits indi­vi­du­als & socie­ty. In Ger­ma­ny, the Ethics Coun­cil made important recom­men­da­ti­ons in 2023. It empha­si­zed that for the ethi­cal eva­lua­ti­on of AI, not only the tech­no­lo­gy must be under­s­tood, but also the inter­ac­tions bet­ween humans and tech­no­lo­gy as well as social effects. Some uni­ver­si­ties alre­a­dy have their own insti­tu­tes for the ethics of AI, e.g. the Tech­ni­cal Uni­ver­si­ty of Munich and the Uni­ver­si­ty of Oxford.

We have the Natio­nal Ethics Com­mit­tee for Human Medi­ci­ne (NEK) and the Ethics Com­mit­tee for Bio­tech­no­lo­gy in the Non-Human Field. It is time to crea­te a third ethics com­mit­tee for chal­lenges in con­nec­tion with AI, its deve­lo­p­ment and use. Ana­log­ous to the NEK, this should be a be an inde­pen­dent body that ana­ly­zes ethi­cal issues and with a view to Switzerland’s demo­cra­tic tra­di­ti­ons, civil liber­ties and fun­da­men­tal social rights. The Com­mis­si­on advi­ses poli­cy­ma­kers, but does not anti­ci­pa­te legis­la­ti­ve decisions.

State­ment of the Fede­ral Coun­cil of 21.5.25

The Fede­ral Coun­cil is awa­re that the deve­lo­p­ment of arti­fi­ci­al intel­li­gence (AI) brings with it new chal­lenges, par­ti­cu­lar­ly at a legal, social and ethi­cal level. The Fede­ral Admi­ni­stra­ti­on alre­a­dy addres­sed ethi­cal issues in its initi­al work in the field of AI, name­ly in the 2019 report on the chal­lenges of AI and in the 2020 “Arti­fi­ci­al intel­li­gence” gui­de­lines for the Con­fe­de­ra­ti­on (available at https://www.sbfi.admin.ch > ERI poli­cy > Fede­ral government’s edu­ca­ti­on, rese­arch and inno­va­ti­on poli­cy 2025 – 2028 > Trans­ver­sal the­mes in the ERI sec­tor > Digi­ta­lizati­on in the ERI sec­tor > Arti­fi­ci­al intel­li­gence). Ethi­cal chal­lenges are also a cen­tral ele­ment of con­side­ra­ti­ons on the gover­nan­ce of AI at inter­na­tio­nal level. On Novem­ber 23, 2021, all mem­ber sta­tes of UNESCO, inclu­ding Switz­er­land, adopted a recom­men­da­ti­on on the ethics of AI. In addi­ti­on, num­e­rous exi­sting inter­na­tio­nal instru­ments on AI, inclu­ding tho­se deal­ing with ethi­cal issues, were taken into account during the nego­tia­ti­ons on the draf­ting of the Coun­cil of Euro­pe Con­ven­ti­on on Arti­fi­ci­al Intel­li­gence and Human Rights, Demo­cra­cy and the Rule of Law (see expl­ana­to­ry report on the AI Con­ven­ti­on, point 7, available at https://www.coe.int > Droits humains > Intel­li­gence arti­fi­ci­el­le et droits humains > Con­ven­ti­on-cad­re).

On Febru­ary 12, 2025, the Fede­ral Coun­cil made a fun­da­men­tal decis­i­on on the regu­la­ti­on of AI in Switz­er­land. It would like to Rati­fy the Coun­cil of Europe’s AI Con­ven­ti­on and has ins­truc­ted the FDJP to do so, a con­sul­ta­ti­on draft by the end of 2026 on their imple­men­ta­ti­on in Swiss law (see press release at https://news.admin.ch > Press Releases of the Fede­ral Coun­cil > Febru­ary 12, 2025 – AI regu­la­ti­on: Fede­ral Coun­cil wants to rati­fy Coun­cil of Euro­pe con­ven­ti­on). The preli­mi­na­ry draft law will deal in par­ti­cu­lar with the chal­lenges rela­ting to trans­pa­ren­cy, data pro­tec­tion, non-dis­cri­mi­na­ti­on and super­vi­si­on. The con­ven­ti­on gives the con­trac­ting sta­tes lee­way in the choice of mea­su­res to be taken. The pos­si­ble crea­ti­on of an ethics com­mit­tee in the field of AI can be exami­ned as part of this work, which is only just begin­ning. It is not yet pos­si­ble to say what spe­ci­fic mea­su­res will be taken.