Analyst Insight: As artificial intelligence continues to expand across industries, some have struggled with its implementation. When applying this new technology, it’s vital to keep a human-centric approach and not lose sight of what keeps supply chains and logistics running smoothly. 

As we continue to navigate through these turbulent times, there is a growing fear from the public about the repercussions of AI. Fifty percent of Americans are more concerned than excited about the technology, and only 10% are more excited than not, according to a Pew Research poll. With this hesitancy comes resistance – public perception impacts how workers approach AI and by proxy, AI’s effectiveness. It then falls on companies to ensure the efficacy and ethics of the AI’s use as these concerns grow.

With AI initially proposed as a tool with endless potential for growth and efficiency, corporations began tripping over themselves to implement these new tools. However, in this mad dash, they began investing without meaningful direction or clearly defined business goals for the solutions. As a result, 95% of trials of generative AI have yielded little ROI, according to an MIT study.

What the remaining 5% of successful pilots uncovered was that AI worked best when focused on decision support and intelligence rather than total autonomy. Data from ABI Research confirms that 94% of companies plan to use AI specifically to assist with decision-making, showing a growing trust in AI-driven recommendations rather than fully autonomous agents. 

This trend validates the need for a “human-centric” approach, harnessing AI to enhance the human laborer by providing the insights necessary to make complex choices, rather than attempting to replace them entirely. It’s about creating “positive friction,” where human expertise guides and validates the supply chain's output.

AI systems are great tools for repetitive tasks, but eliminating every source of oversight can have consequences. When approaching the data, some may be led to dismiss AI as just another fad. But like any other tool, when AI is introduced into the workforce, time must be taken to analyze and assess how to best implement it in conjunction with existing operations.

When AI is divorced from oversight or input, operations risk “double work,” or deliverables that must be re-done due to errors present from AI output. When AI operates without accountability, it can generate unforeseen mistakes and require employees to go back and correct them or override them, decreasing efficiency and further demoralizing employees by expediting burnout if teams were downsized to accommodate these AI systems.

The key is keeping the “human” at the center. When approaching these new tools, especially within the complexities of vehicle logistics, AI should be symbiotic in nature through “tool interaction” and an “observation-thought-action” loop, working autonomously, but not independent of human control or oversight.

Supply chains and logistics are made up of many complex and intertwined layers that require both a basic level of knowledge and the ability to anticipate and predict needs in the future. This “foresight” is something that AI can often lack as it is only capable of acting on the data it is trained on.

Tool interaction means allowing seamless communication between the AI tools and existing databases and software to create a unified working experience and to mitigate gaps in operations.

By allowing the worker to set the goals and then allowing AI to act on them, workers can attend to more immediate and complex concerns instead of the more repetitive tasks while not fully removing them from the process.

One of the keys to maintaining a strong ethical operation is maintaining transparency. AI is most unstable when there’s a lack of understanding either by the user, the one implementing it, or both. At all levels of the supply chain, individuals should be able to understand the purpose, both short and long term, and the results of any tool.

To accomplish this, AI should abide by four core elements.

• Beneficial. Having a clear benefit to users and society while also mitigating negative factors like bias or misinformation.
• Aligned. Ensuring AI is in line with the values of a business, with understandable AI systems as its foundation.
• Reliable. Prioritizing quality, transparency, and consistency within these systems.
• Privacy-protected. Upholding a standard of privacy that doesn’t infringe or overbear on the individual or break existing laws and regulations.

These basic guidelines will offer confidence in AI to stakeholders at every level. Workers feel the benefit of independent systems without losing their own autonomy, and companies can increase their own efficiency, ROI and output thanks to staff at all levels gaining more bandwidth and visibility.