An article for business leaders, intelligence professionals, and decision-makers navigating the challenges of integrating AI into critical workflows.

What you will learn: Explore how to differentiate between reliable and less trustworthy AI models for business decisions. Understand the risks of relying on consumer AI platforms and learn strategies to implement trustworthy AI systems that balance machine intelligence with human judgement.

Trust in AI: Balancing Confidence and Caution

Eight in 10 (81%) businesses believe that generative AI provides accurate and relevant intelligence, with 86% deeming these platforms to be altogether trustworthy. This is a very promising result amidst the diverse and sometimes contradictory perceptions among professionals.

Artificial intelligence has become integral to the business decision-making process, offering unprecedented capabilities to analyse and interpret vast datasets with speed and precision. However, a paradox emerges from AMPLYFI’s data: while many praise AI for its contributions to business intelligence, there is a notable hesitance about its reliability.

Over three quarters (78%) of knowledge workers express growing concerns that popular generative AI models like ChatGPT are eroding people’s trust in AI.

This raises a critical question: how can businesses make sure they are trusting the right type of AI for important decisions?

Discover the data behind our insights

The challenges of popular AI models

Recent user complaints over ChatGPT’s ‘lazy’ behaviour have become increasingly common, with the makers of ChatGPT themselves acknowledging that “model behaviour can be unpredictable” sometimes. Headlines have emphasised this by focusing on the inaccuracies and limitations of AI models like ChatGPT, overshadowing the potential of more sophisticated AI systems.

Examples of perceived laziness in consumer AI have included limited response lengths, sometimes dubbed as ‘quiet quitting’ in AI, and mimicking human productivity patterns, referred to as the ‘winter break hypothesis’, which suggests that AI might vary its productivity similar to human slowdowns during less active periods.

There have also been cases of what is referred to as “AI hallucinations”. These are the generation of incorrect or misleading results due to various factors, including insufficient training data, biases in training data, and incorrect model assumptions.

These inconsistent behaviours in consumer AI platforms like ChatGPT raise valid concerns about their reliability, especially when users expect comprehensive and thorough responses. Such shortcomings could mistakenly be perceived as deceitful behaviour by AI.

Building trustworthy AI solutions

Navigating the challenges posed by AI inaccuracies is crucial for ensuring effective decision-making in the business landscape. Here are some strategic measures businesses can adopt to mitigate the risks and enhance the reliability of AI applications, including:

Navigating the future with confidence

Platforms powered by trustworthy AI go beyond the surface-level information that dominates conventional search engine results without the risk of artificial hallucinations and inconsistent answers. This ultimately is the cause for most people’s distrust. Unlike some models that offer accessible but unreliable answers, trustworthy AI prioritises verified evidence and in-depth analysis.

Businesses should therefore adopt AI with strategic safeguards to ensure that their reliance on this technology is informed, measured, and proactive, thus maximising AI’s potential benefits while minimising risks.

For a deeper dive into the mechanics of trust and how to leverage AI effectively and reliably in your organisation, explore AMPLYFI’s comprehensive report: