Many businesses still don't trust their AI systems - and that could be a major problem

• Businesses don't trust the accuracy of their AI/ML models, but it's due to poor data foundations, report claims
• Only one in three have implemented or optimized data observability programs
• Observability should be standard across the whole data lifecycle

New research from Ataccama has claimed a considerable proportion of businesses still don't trust the output of AI models - but this could simply be because their data isn't in order yet.

The study found two in five (42%) organizations don't trust their AI/ML model outputs, yet only three in five (58%) have implemented or optimised data observability programs.

Ataccama says this could be a problem, because traditional observability tools are not designed to monitor unstructured data, such as PDFs and images.

Don't trust AI? A lack of suitable data could be the problem

The report also revealed the ad-hoc approach that businesses often take, with observability often implemented reactively, resulting in fragmented governance and silos across the organization.

Ataccama defined an effective program as proactive, automated and embedded across the data lifecycle. More advanced observability could also include automated data quality checks and remediation workflows, which could ultimately prevent further issues upstream.

"They’ve invested in tools, but they haven’t operationalized trust. That means embedding observability into the full data lifecycle, from ingestion and pipeline execution to AI-driven consumption, so issues can surface and be resolved before they reach production," CPO Jay Limburn explained.

However, ongoing skills shortages and limited budgets are still presenting challenges along the way. Ataccama also noted that unstructured inputs continue to grow as a result of increased generative AI and RAG adoption, yet currently, fewer than one in three organizations feed unstructured data into their models.

The report goes on to explain: "The most mature programs are closing that gap by integrating observability directly into their data engineering and governance frameworks."

With proper observability in place, businesses can expect improved data reliability, faster decision-making and reduced operational risk.

You might also like

• These are the best AI tools and best AI writers
• Many businesses are thinking twice on using AI bots
• Check out our roundup of all the best cloud hosting providers