Trust in Artificial Intelligence

As featured on PhilStar:  Trust in Artificial Intelligence

Artificial Intelligence (AI) has become a ubiquitous part of everyday life and work. AI is transforming the way work is done and how services are delivered. Given its potential and realized benefits for organizations, employees, and society, interest in AI continues to grow. Organizations are leveraging the remarkable power of AI to help improve data-based predictions, optimize products and services, augment innovation, enhance productivity and efficiency, and lower costs.

However, AI adoption also poses risks and challenges, raising concerns about whether AI use today is truly trustworthy. These concerns have been fueled by high-profile cases of AI use that were biased, discriminatory, manipulative, unlawful or in violation of human rights. Realizing the potential benefits of AI, and a return on investment, requires a clear and sustained focus on maintaining the public’s trust. To drive adoption, people need to be confident that AI is being developed and used in a responsible and trustworthy manner.

In collaboration with the University of Queensland, KPMG Australia led the world-first deep dive into trust and global attitudes towards AI across 17 countries. Trust in artificial intelligence: A global study 2023 provides broad-ranging global insights into the drivers of trust, the perceived risks and benefits of AI use, community expectations of governance of AI, and who is trusted to develop, use and govern AI.

Most people are wary about trusting AI systems and have low or moderate acceptance of AI. Trust and acceptance depend on the AI application.

• Three in five (61 percent) are wary about trusting AI systems.
• 67 percent report low to moderate acceptance of AI.
• AI use in human resources is the least trusted and accepted, while AI use in healthcare is the most trusted and accepted.
• People in emerging economies are more trusting, accepting and positive about AI than people in other countries.

People recognize AI’s many benefits, but only half believe the benefits outweigh the risks. People perceive AI risks in a similar way across countries, with cybersecurity rated as the top risk globally.

• 85 percent believe AI results in a range of benefits.
• Yet only half of respondents believe the benefits of AI outweigh the risks.
• Top concern is cybersecurity risk at 84 percent.

People are most confident in universities and defense organizations to develop, use and govern AI and they are least confident in government and commercial organizations.

• 76 to 82 percent confidence in national universities, research institutions and defense organizations to develop, use and govern AI in the best interest of the public.
• One-third of respondents lack confidence in government and commercial organizations to develop, use and govern AI.

There is strong global endorsement for principles that define trustworthy AI. Trust is contingent on assuring such principles are in place. People expect AI to be regulated with external, independent oversight — and they view current regulations and safeguards as inadequate.

• 97 percent strongly endorse the principles for trustworthy AI.
• Three in four would be more willing to trust an AI system when assurance mechanisms are in place.
• 71 percent expect AI to be regulated.

Most people are comfortable using AI to augment work and inform managerial decision-making but want humans to retain control.

• About half are willing to trust AI at work.
• Most people are uncomfortable with or unsure about AI use for HR and people management.
• Two in five believe AI will replace jobs in their area of work.
• Younger people, the university educated and managers are more trusting of AI at work.

People want to learn more about AI but currently have a low understanding. Those who understand AI better are more likely to trust it and perceive greater benefits

• Half of respondents feel they don’t understand AI or when and how it’s used.
• 45 percent don’t know AI is used in social media.
• 85 percent want to know more about AI.

However, just like in many other countries, industry leaders have grown wary of trusting AI systems. Therefore, they are obligated to establish robust AI governance frameworks that can inspire confidence among the public.

“Building trust in AI systems underscores the need for responsible AI development, governance and continuous evaluation. This ensures that AI aligns with societal expectations and remains a positive force for progress in the Philippines,” Royeca added.

Trust is central to the acceptance of AI and is influenced by four key drivers.

The survey analysis demonstrates that trust is critical to AI acceptance and adoption. Through modeling, four distinct pathways to trust have been identified, representing key drivers that influence trust in AI and that is expected to strengthen responsible AI use. The pathways are institutional, motivational, uncertainty reduction and knowledge. Of these significant and complementary drivers, the institutional pathway had the strongest influence on trust, followed by the motivational pathway.

1.     The institutional pathway

Reflects the beliefs about the adequacy of current safeguards, regulations and laws to make AI use safe, and confidence in government and commercial organizations to develop, use and govern AI.

2.     The motivational pathway

Reflects the need to demonstrate the potential benefits of AI use to motivate trust.

3.     The uncertainty-reduction pathway

Reflects the need to address concerns about AI risks.

4.     The knowledge pathway

Reflects people’s understanding of AI use and efficacy in using technology.

These insights are relevant for informing responsible AI strategies, practices and policies within businesses, government and NGOs at a national level, as well as informing AI guidelines, standards and policy at the international and pan-governmental levels.

There’s a range of resources to support organizations in embedding principles and practices of trustworthy AI into their operations and putting in place mechanisms that support stakeholder trust in the use of AI. While proactively investing in these trust foundations can be time and resource intensive, this research suggests it’s critical for sustained acceptance and adoption of smart technologies over time and hence a return on investment.

Given AI’s rapid and widespread deployment, it is expected to be important to regularly re-examine public trust and expectations of AI systems as they evolve over time to help ensure AI use is aligned with and meeting changing public expectations.

_{The excerpt was taken from the KPMG Thought Leadership publication: https://kpmg.com/xx/en/home/insights/2023/09/trust-in-artificial-intelligence.html?cid=linkd-soc_linkd_2023_sep_21_trust_ai__1_social}