Robot-psy­chol­o­gist Mar­ti­na Mara and AI-expert Sepp Hochre­it­er dis­cussed with Markus Hengstschläger and stu­dents in the con­text of the Euro­pean Forum Alp­bach, whether mankind should rely on arti­fi­cial­ly intel­li­gent systems.

Robots must still earn trust

Smart tech­nolo­gies con­trol or guide more and more areas of human life are. How­ev­er, Robots and arti­fi­cial intel­li­gent sys­tems still have to work hard to earn the trust of humankind. Mar­ti­na Mara (Ars Elec­tron­i­ca Future­lab) and Sepp Hochre­it­er (Insti­tute of Bioin­for­mat­ics of the JKU-Linz) agreed on that at the fire­side-talk orga­nized by Club Alp­bach Oberöster­re­ich togeth­er with ACADEMIA SUPERIOR.

Artificial intelligence will change everything

The rel­e­vance of the top­ic is high. Pres­i­dent Michael Strugl was con­vinced: „Arti­fi­cial intel­li­gence will change every­thing”. And Hochre­it­er, who will lead the new Arti­fi­cial Intel­li­gence Lab at the JKU, pre­dict­ed, „In ten years, we will car­ry the brain capac­i­ty of a human being with us in every smart­phone. Then, arti­fi­cial intel­li­gence will have more impact on mankind than the inven­tion of the computer”.

Hochre­it­er is a research-pio­neer on arti­fi­cial intel­li­gence and devel­oped algo­rithms for neur­al net­works and deep learn­ing already 20 years ago. His sci­en­tif­ic work is now found in almost every smart­phone and is basis, for exam­ple, of the voice recog­ni­tion tech­nol­o­gy on mobile devices. The whole IT-world is fol­low­ing to his research at the JKU-Linz.

Making robots predictable

For that research can be car­ried out in real­i­ty, peo­ple have to rely more on arti­fi­cial­ly intel­li­gent sys­tems. The research-focus of Mar­ti­na Mara. She deals with how robots should look, com­mu­ni­cate, and behave, so that peo­ple feel com­fort­able with them. „The pre­dictabil­i­ty of robot behav­ior is cru­cial for the intro­duc­tion of arti­fi­cial­ly intel­li­gent sys­tems,” told Mara about the expe­ri­ence of auto­mo­tive indus­try in the design of self-dri­ving cars.

„Intel­li­gent sys­tems must be able to explain why they have made what kind of deci­sion.” — Mar­ti­na Mara

Tech­ni­cal­ly, accord­ing to Mara, is today more pos­si­ble than is imple­ment­ed in the mar­ket. What hin­ders autonomous cars in the mar­ket are main­ly eth­i­cal, moral and legal issues, but also the skep­ti­cism of the con­sumers. Arti­fi­cial intel­li­gence is not only because of films like Ter­mi­na­tor or Blade Run­ner, in which machines turn against humans, con­nect­ed with fears. „An intel­li­gent sys­tem would nev­er make some­thing stu­pid like that,” Hochre­it­er plead­ed passionately.

How to design smart systems smart

It is cru­cial to start with the design in order to counter these fears. „Machines should be designed in such a way that they are imme­di­ate­ly rec­og­niz­able as machines,” Mara referred to a phe­nom­e­non known as „uncan­ny val­ley” or „accep­tance gap”: human-like machines do not pro­duce trust, but fear. „As soon as robots claim they are human, it becomes creepy. The clos­er robots get to us, the more uncom­fort­able it becomes,” says Mara, who is a mem­ber of the new­ly found­ed Aus­tri­an Coun­cil for Robots, which will devel­op a local strat­e­gy for the han­dling of intel­li­gent machines.

„Through arti­fi­cial intel­li­gence, peo­ple will not speak less, but more with each oth­er.” — Sepp Hochreiter

The live­ly dis­cus­sion with the stu­dents raised ques­tions such as whether robots will trig­ger lone­li­ness or depres­sion in peo­ple, how the coop­er­a­tion between man and machine will be designed in the future and when self-pro­pelled cars can be expect­ed in the mass mar­ket. In 2024, Hochre­it­er and Mara agreed. But first of all, such cars might still have to enter the „car kinder­garten” to get a driver’s license.

Among the 70 dis­cus­sion guests were among oth­ers the stu­dents and schol­ar­ship hold­ers of the Club Alp­bach Oberöster­re­ich. Also Rec­tor Univ.-Prof. Dr. Mein­hard Lukas (JKU-Linz), Eve­line Pupeter (Empo­ria), DI (FH) Wern­er Pam­minger (Eco­nom­ic Agency of Upper Austria).