As artificial intelligence (AI) spreads across supply chain and logistics operations, growing public anxiety toward the technology is directly shaping enterprise adoption and ROI outcomes. According to Pew Research, 50% of Americans are more concerned than excited about AI, with only 10% more excited than not. This hesitancy ripples through the workforce — workers' attitudes toward AI shape its effectiveness — placing the burden squarely on companies to ensure both ethical and effective use of the technology.
Initial enterprise AI spend, framed as a tool with limitless growth potential, has been heavily eroded by deployments without meaningful direction or clearly defined business goals. An MIT study finds that 95% of generative-AI pilots have produced little ROI. The successful 5% share a common pattern: they position AI as decision support and intelligence rather than full autonomy. ABI Research confirms the direction of travel — 94% of companies plan to use AI specifically to assist with decision-making, signaling growing trust in AI-driven recommendations rather than fully autonomous agents.
This validates a "human-centric" design principle that uses AI to enhance the human operator with insights for complex choices, rather than to replace them. The pattern is sometimes called "positive friction": human expertise guides and validates the AI's output. AI is excellent for repetitive tasks, but eliminating every source of oversight risks "double work" as people re-do AI errors, accelerating burnout if teams have been downsized to absorb the AI. The recommended architecture is built on tool interaction and an "observation-thought-action" loop — AI communicates seamlessly with existing databases and software, workers set the goals, and AI acts on them while humans retain oversight.
For supply chain operations, human-centric AI design should rest on four pillars: beneficial, with clear value to users and society while mitigating bias and misinformation; aligned, with understandable AI systems matching the values of the business; reliable, prioritizing quality, transparency and consistency; and privacy-protected, respecting individuals and existing regulatory frameworks. AI built on these pillars lets workers gain the leverage of independent systems without giving up their own autonomy, while companies see efficiency, ROI and visibility improvements — especially in layered processes like vehicle logistics, where AI's "foresight" gap (limited to its training data) becomes manageable only when paired with human judgement on the ground.
Key Takeaways:
1. Pew Research finds 50% of Americans are more concerned than excited about AI; the hesitancy directly shapes enterprise adoption and ROI outcomes.
2. An MIT study finds 95% of generative-AI pilots produced little ROI; the successful 5% positioned AI as decision support, not full autonomy.
3. ABI Research reports 94% of companies plan to use AI specifically to assist with decision-making.
4. Human-centric AI uses positive friction, an observation-thought-action loop and tool interaction to keep humans in the loop; unsupervised AI creates double work and burnout.
5. Supply-chain AI should rest on four pillars: beneficial, aligned, reliable and privacy-protected.
[43966]
