Urgent.News

600+ sources. One page. See who else covered it.

Editions

AI

Is AI really helping your SMB? Study finds a quarter of execs can't explain what their AI actually does

An incredible 25% of business leaders are unable to explain AI-generated outputs, while a worrying majority rely on AI for vital financial tasks, including expenses and payments

Is AI really helping your SMB? Study finds a quarter of execs can't explain what their AI actually does

A recent study has revealed that a quarter of business executives struggle to explain the workings of the AI systems they employ, potentially putting their small and medium-sized enterprises (SMEs) at risk. The survey, conducted by Startup.co.uk, found that 25% of business leaders find it difficult, and 12% find it very difficult, to articulate how AI-generated outputs are derived.

This lack of understanding raises concerns about the trust SMEs place in AI for complex financial tasks, including audits, compliance, and financial management.

Customers and investors are increasingly wary of businesses that rely on unverified AI, as they may perceive it as a risk to their data and security. The survey also highlighted that 85% of small businesses are using AI for financial tasks, with 37% employing AI for accounts payable processes, 32% for audit and compliance, and 31% for managing expenses.

While AI can offer benefits such as fraud detection and performance insights, the inability of leaders to explain AI logic could lead to significant legal and compliance issues, such as breaching GDPR regulations.

With fines of up to £17.5 million or 4% of global turnover for verified GDPR breaches, businesses must ensure they understand the processes behind AI systems. The study suggests that blindly trusting AI with sensitive financial data is a risky gamble, as the implications of data misuse could be severe. Therefore, business leaders need to improve their understanding and communication of AI's functionality to mitigate potential risks and maintain compliance with data protection laws.

Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at techradar.com →

More in AI

From Midnight Power Cuts to Multi-Agent Voice AI: How I Built Raksha in 10 Days

Building voice AI sounds deceptively clean on paper: capture speech, stream it to an STT engine, prompt an LLM, and synthesize audio back in real time.

  • Class 12 student built voice assistant Raksha for #VoiceForBharat Challenge
  • Raksha protects citizens from cyber scams and verifies government schemes
  • Developed in 10 days despite power cuts and technical challenges

An AI Capture-the-Flag Tournament: What the Scoreboard Counted

Code: Megapixel99/capture-the-flag In April I ran five games of an AI capture-the-flag tournament between five small open-weight models (1.0B to 2.5B parameters).

  • Qwen 3.5 captured 13 flags without losing any in AI capture-the-flag tournament
  • Larger models showed no confirmed impact on security reasoning or multi-step exploitation
  • RNJ-1 8B finished last despite having more parameters than other models

More from Saturday 15 August →