Í alþjóðlegri rannsókn KPMG er fjallað um traust fólks á gervigreind, ábyrga gervigreind, væntingar til stjórnunar gervigreindar og ávinning og áhættu fyrir fyrirtæki og samfélag.

Í samstarfi við háskólann í Queensland vann KPMG í Ástralíu fyrstu alþjóðlegu rannsóknina á hnattrænum viðhorfum til gervigreindar (AI).

Gervigreind er að breyta daglegu lífi með tilkomu spjallmenna, samfélagsmiðla, andlitsþekkingar, forvörnum gegn svikum og svo mætti lengi telja en þrátt fyrir mikla útbreiðslu er víða ekki nægur skilningur á gervigreind.  

Í könnuninni er viðhorf og traust almennings til gervigreindar í 17 löndum rannsökuð og ljósi varpað á helstu drifkrafta trausts, væntingar samfélagsins um stjórnun og stjórnun gervigreindar, hvernig fólk er að upplifa ávinning og áhættu af gervigreind og hvað fólki finnst um notkun hennar í vinnu. 

Helstu niðurstöður rannsóknarinnar á ensku

Lykilniðurstöður rannsóknarinnar innihalda:

1. AI trust and acceptance

Three out of five people (61 percent) are either ambivalent or unwilling to trust AI.

But younger generations and the university educated are more accepting. People have more faith in the capability and helpfulness of AI systems, and are more sceptical of their safety, security and fairness.

2. How well do people understand AI?

While most people (82 percent) have heard of AI, about half (49 percent) are unclear about how and when it is being used. However, most (82 percent) want to learn more.

What’s more, 68 percent of people report using common AI applications, but 41 percent are unaware AI is a key component in those applications.

3. Benefits of AI

Most people (85 percent) believe AI will deliver a range of benefits, including efficiency, effectiveness, innovation and resources.

On average, only one in two people believe the benefits of AI outweigh the risks. People in India, China, South Africa and Brazil have the most positive views of the benefits of AI. Conversely, Australia, Canada, UK and Japan are the least positive.

4. Risks of AI

Alongside the benefits, most people (73 percent) perceive significant risks in AI.

Cyber security is rated as the top risk globally, followed by harmful use of AI, job loss (especially in India and South Africa), loss of privacy, system failure (particularly in Japan), deskilling, and undermining human rights.

5. Expectations around AI regulation

Most people (71 percent) believe AI regulation is necessary.

Many expect some form of external, independent oversight and only 39 percent believe current governance and regulations are enough to protect people and make AI use safe.

6. Principles of responsible AI

There is strong global endorsement for the principles of trustworthy AI originally proposed by the European Union, with almost everyone (97 percent) viewing these principles as important for trust.

7. AI in the workplace

One-third of people believe they never use AI in the workplace, and a further 13 percent are unsure. This suggests many don’t fully understand where and how AI is currently used.

Most people (55 percent) are comfortable with the use of AI at work to augment and automate tasks in managerial decision-making. People actually prefer AI involvement to sole human decision-making, but they want humans to retain control.

Except in China and India, most people believe AI will remove more jobs than it creates.

8. Shifting attitudes

We surveyed attitudes towards AI in Australia, the UK, USA, Canada, and Germany in 2020.

Trust in AI has increased in each of these countries from 2020 to 2022.

More people have heard of AI in 2022 (78 percent) than in 2020 (62 percent) and are more likely to identify AI use in common applications, but there was no change in subjective understanding.

There was no change in the perceived adequacy of AI governance and regulation, nor confidence in entities to develop, use and govern AI.