Varna in učinkovita uporaba umetne inteligence pri delu

(Slovenska verzija)

Umetna inteligenca lahko seveda ponudi korist podjetjem, vendar prinaša tudi tveganja. Kako dobro vi upravljate s temi tveganji? Če imate opravka z občutljivimi podatki ali vas potencialne napake lahko drago stanejo, je zelo pomembno, da z novimi orodji umetne inteligence delate varno in učinkovito.

Prav to je bila tema treninga “Varna in učinkivita uporaba umetne inteligence pri delu”, ki sem ga vceraj izvedel za zaposlene pri iConsult d.o.o, svetovanje za davčne, pravne in finančne zadeve. Ogledali smo

• Kako deluje umetna inteligenca,
• Kako predvideti, kdaj in kako vam lahko pomaga, česa nikakor ne zmore, ter kje se skrivajo pasti njene uporabe,
• Kaj so občutljivi podatki, kaj so osebni podatki, ter kaj predpisi določajo glede načina obdelave podatkov.

Človeska narava, še posebej naša radovednost, pomenita, da ni dovolj, da se pri treningu osredotočamo samo na to, čemu se izogibati ali kakšne morajo biti varovalke. Podjetja potrebujejo tudi pozitivne primere, kako lahko zaposleni varno in koristno uporabljajo umetno inteligenco.

Zato smo si na koncu treninga ogledali tudi, kako je mogoče združiti in nadgraditi prizadevanja posameznikov v podjetju, ki so se najprej lotili uporabe umetne inteligence, s prednostmi pravilnega ravnanja s podatki na nacin, da je to v korist celotnega podjetja.

(English version)

AI advances offer companies many potential advantages, but how can they reliably manage the accompanying risks? Managing AI risks is especially important for companies dealing with confidential data, and for which AI errors are costly. Companies like iConsult d.o.o, experts in tax, legal and financial topics where yesterday I led a training “Safe and Effective AI at Work.”

We started by covering

• How AI works, how it fails, and how to predict which will happen when.
• What is sensitive data, what is personal data, and what do EU and national laws require about how businesses process them.

It’s never enough, however, to only tell employees what not to do in a training. Human nature, especially our curiosity, means that companies need to offer its employees positive options for safe AI usage.

The final section of the training focused precisely on examples of this: how the individual work of a company’s early AI adopters can be brought together to the benefit of the entire company. This means harnessing as a company the benefits of proper data handling while also collectively benefitting from the early adopters’ tips and tricks for getting reliable, safe benefit from AI tooling at work.