• Home 
  • -
  • AI 
  • -
  • Inter­pel­la­ti­on Cot­tier (23.3930): Arti­fi­ci­al intel­li­gence. What frame­work con­di­ti­ons must be crea­ted to make the most of it and avo­id unde­si­ra­ble developments? 

Inter­pel­la­ti­on Cot­tier (23.3930): Arti­fi­ci­al intel­li­gence. What frame­work con­di­ti­ons must be crea­ted to make the most of it and avo­id unde­si­ra­ble developments?

Inter­pel­la­ti­on Cot­tier (23.3930): Arti­fi­ci­al intel­li­gence. What frame­work con­di­ti­ons must be crea­ted to make the most of it and avo­id unde­si­ra­ble developments?

Sub­mit­ted text

Arti­fi­ci­al intel­li­gence (AI) is rapid­ly play­ing an incre­a­sing­ly important role in our socie­ty. The poten­ti­al for pro­gress offe­red by the­se tech­no­lo­gies is immense. So are the chal­lenges they bring with them.

AI has the poten­ti­al to radi­cal­ly chan­ge the way we work, com­mu­ni­ca­te and even live tog­e­ther. If it is used incor­rect­ly or for the wrong pur­po­ses, it can also have serious nega­ti­ve con­se­quen­ces, par­ti­cu­lar­ly for pri­va­cy, secu­ri­ty or human dignity.

In this con­text, I would like to put the fol­lo­wing que­sti­ons to the Fede­ral Council:

  • Is he rea­dy to launch a natio­nal rese­arch pro­gram that looks at how AI will affect our socie­ty and demo­cra­cy, what frame­work con­di­ti­ons can be crea­ted for it and how it can be deve­lo­ped for pur­po­ses that bene­fit everyone?
  • Does it plan to crea­te a spe­ci­fic legal frame­work for tech­no­lo­gies that use AI, or does it con­sider the cur­rent legis­la­ti­on to be sufficient?
  • How does he think the new EU regu­la­ti­on will affect Swiss legis­la­ti­on? Will Switz­er­land have to take mea­su­res to adapt or to obtain reco­gni­ti­on of equivalences?
  • What impact can the ongo­ing nego­tia­ti­ons on a Coun­cil of Euro­pe Con­ven­ti­on on “Arti­fi­ci­al Intel­li­gence, Human Rights, Demo­cra­cy and the Rule of Law” have on Switz­er­land and its legis­la­ti­on? The preli­mi­na­ry draft of the con­ven­ti­on was recent­ly published by the Com­mit­tee on Arti­fi­ci­al Intel­li­gence (CAI), which is chai­red by a Swiss natio­nal. Is Switz­er­land plan­ning to adapt its legis­la­ti­on? Has work alre­a­dy begun in this area?
  • What mea­su­res should be taken to ensu­re that the deve­lo­p­ment of the­se tech­no­lo­gies finds a regu­la­ted but favorable frame­work in Switz­er­land and that com­pa­nies acti­ve in this field are foun­ded or set up in Switzerland?
  • Should­n’t Switz­er­land deve­lop a com­pre­hen­si­ve stra­tegy that covers the­se dif­fe­rent aspects? Is the Fede­ral Coun­cil pre­pared to tack­le this task?

Opi­ni­on of the Fede­ral Coun­cil of 30.8.2023

The Fede­ral Coun­cil has alre­a­dy taken num­e­rous mea­su­res to ensu­re the respon­si­ble use of arti­fi­ci­al intel­li­gence (AI). On the basis of the report published in 2019 by the inter­de­part­ment­al working group on arti­fi­ci­al intel­li­gence, it has drawn up gui­de­lines for the use of AI in the Fede­ral Admi­ni­stra­ti­on and has published the Com­pe­tence Net­work for Arti­fi­ci­al Intel­li­gence (CNAI) was launched.

The­re are estab­lished pro­ce­du­res for laun­ching a natio­nal rese­arch pro­gram. Inte­re­sted par­ties can sub­mit topic pro­po­sals accor­ding to the bot­tom-up prin­ci­ple. Based on the­se, the Sta­te Secre­ta­ri­at for Edu­ca­ti­on, Rese­arch and Inno­va­ti­on (SERI) draws up pro­gram pro­po­sals, which are checked for fea­si­bi­li­ty by the Swiss Natio­nal Sci­ence Foun­da­ti­on. The Fede­ral Coun­cil has alre­a­dy laun­ched the Natio­nal Rese­arch Pro­gram NRP 77 “Digi­tal Trans­for­ma­ti­on”, which also inclu­des rese­arch on AI – par­ti­cu­lar­ly in the rese­arch modu­le “Ethics, Trust­wort­hi­ness and Gover­nan­ce”. A pro­po­sal on digi­tal demo­cra­cy was also exami­ned in the NRP 2022/2023 review round. This review was nega­ti­ve, part­ly becau­se many over­laps with NRP 77 were iden­ti­fi­ed (see respon­ses to Inquiry 23.7210 and Inter­pel­la­ti­on 23.3281).

The effects of the use of AI are high­ly con­text-depen­dent. For this rea­son, the Fede­ral Coun­cil has alre­a­dy taken regu­la­to­ry mea­su­res for cer­tain are­as. In par­ti­cu­lar, it has ins­truc­ted the Fede­ral Depart­ment of the Envi­ron­ment, Trans­port, Ener­gy and Com­mu­ni­ca­ti­ons (DETEC) to prepa­re a draft on the regu­la­ti­on of lar­ge com­mu­ni­ca­ti­on plat­forms. This draft is inten­ded to streng­then the rights of users in Switz­er­land and increa­se the trans­pa­ren­cy requi­re­ments for plat­forms. The new Data Pro­tec­tion Act, which comes into force on Sep­tem­ber 1, 2023, also aims to adapt data pro­tec­tion to the sta­te of the art. It con­ta­ins pro­vi­si­ons on decis­i­ons based exclu­si­ve­ly on auto­ma­ted data pro­ce­s­sing. As part of the cur­rent con­sul­ta­ti­on on the amend­ment to the Copy­right Act (ancil­la­ry copy­right for media publishers), it is to be deter­mi­ned whe­ther the­re is a need for regu­la­ti­on with regard to the use of gene­ra­ti­ve AI.

At an inter­na­tio­nal level, the Fede­ral Coun­cil is fol­lo­wing the inter­nal EU nego­tia­ti­ons on the regu­la­ti­on estab­li­shing har­mo­ni­zed rules for arti­fi­ci­al intel­li­gence (Arti­fi­ci­al Intel­li­gence Act, “AI Act”). The­se nego­tia­ti­ons are expec­ted to be con­clu­ded at the end of 2023 or begin­ning of 2024. Befo­re a pos­si­ble adap­t­ati­on of the Swiss legal frame­work to regu­la­te the use of AI in line with the AI Act is con­side­red, the con­se­quen­ces must be ana­ly­zed in detail and the inte­rests careful­ly weig­hed up. Howe­ver, such an ana­ly­sis can only be car­ri­ed out once a final regu­la­ti­on text is available. In addi­ti­on, Switz­er­land is actively invol­ved in the Coun­cil of Europe’s nego­tia­ti­ons on a con­ven­ti­on in this area, con­tri­bu­ting its inte­rests and values. Switz­er­land is chai­ring the rele­vant com­mit­tee. If Switz­er­land rati­fi­es the Con­ven­ti­on, it will then be obli­ged to imple­ment it natio­nal­ly. Whe­ther legal adjust­ments are neces­sa­ry will only beco­me clear once the nego­tia­ti­ons have been concluded.

In its state­ment on the Postu­la­te Dobler (23.3201) the Fede­ral Coun­cil has alre­a­dy sta­ted that it will car­ry out a poli­ti­cal ana­ly­sis within the frame­work of exi­sting fede­ral bodies (in par­ti­cu­lar the inter­de­part­ment­al EU Digi­tal Poli­cy Coor­di­na­ti­on Group, the Tri­par­ti­te Plat­form, the AI Gui­de­lines Moni­to­ring and the FSO’s AI Com­pe­tence Net­work (CNAI)). All fede­ral agen­ci­es respon­si­ble for the legal are­as con­cer­ned are invol­ved in this work. The need for action and pos­si­ble opti­ons for sec­to­ral and, if neces­sa­ry, hori­zon­tal mea­su­res will be iden­ti­fi­ed by the end of 2024.