Founder and Chief Strategy Officer of Arkestro, a leading Predictive Procurement Orchestration platform.
getty
Supply chain and procurement leaders have watched the launch of ChatGPT and other generative AI bots like Bard and Claude with intense excitement and curiosity.
Indeed, generative AI may soon be heavily supporting human decision-making across the enterprise in sales, marketing, purchasing and supply chain. Just consider that IBM announced it will pause hiring to fill 7,800 current and future roles that its CEO believes can be done by AI. However, while there are massive areas of potential, there are equally significant risks posed to privacy, algorithmic bias and even business continuity.
I've been having discussions with other leaders about these issues, especially as they pertain to the safety and ethics policy for governing suppliers' compliant use of AI. Based on these conversations, here are a few of the risk areas that are on leadership teams' minds as well as some commonsense guidelines for procurement teams to ensure that AI operates ethically, accurately and securely.
Risks
While the number of risks that pertain to AI may not yet be fully known, here are a few that are already pertinent to address:
Fraud And Al Spear-Fishing
Security researchers believe that ChatGPT is causing a "tidal wave" of scams. Any scam that targets a supplier, even a small one, can be used to compromise the procurement team of its customers. Consider the 2021 attack involving SolarWinds and other companies for an understanding of how a compromised supplier can affect other companies downstream.
Conversely, suppliers may worry about the accuracy or fairness of an AI selecting the winning supplier as a result of a sourcing process—customers or hackers could potentially manipulate the AI system to favor certain suppliers over others, meaning a potential loss of a lucrative contract.
Intellectual Property Theft
Unless used properly, data entered into ChatGPT or other generative AI tools can spread to other users or companies who are using the service—exposing sensitive IP and "trade secrets." This could lead to serious legal consequences, especially considering laws like the General Data Protection Regulation (GDPR) in the EU, and could harm the company's reputation. In response, some companies like Samsung have already banned or paused the use of generative AI technology by their employees while investigating potential risks.
Algorithmic Bias
Many new AI systems provide algorithmic recommendations for suppliers. Unfortunately, AI are trained on historical data and past decisions, which may include (unconscious) bias against marginalized suppliers. If biases present in the data are not properly addressed, AI could perpetuate or even exacerbate these biases, leading to unfair outcomes and potential legal liability.
We've already seen how these biases have played out in AI that assists in hiring decision-making, which uses similar technology. For example, New York has just adopted a law that will hold companies legally responsible for algorithmic bias in their automated hiring tools.
Supply Disruptions
In my conversations with supply chain leaders, some fear that AI could detect or preempt a run on a specific commodity and trigger an autonomous buying cycle ahead of that run, thereby causing it. It's easy to imagine that as AI becomes more fused with autonomous systems that monitor markets and conduct buy cycles for repetitive transactions, the risk of causing a shortage by trying to "beat the market" ahead of a price spike could unintentionally cause a disruption.
For categories such as food, fuel and medicine, the consequences could prove painful to procurement and supply management teams—not to mention the customers and communities that we serve. AI-managed inventory decisions could also lead to unintended inventory levels—causing little inventory, backorders, increased logistics costs and loss of revenues and customer confidence.
Commonsense Guidelines
Given these threats, and how quickly AI is evolving, it makes sense to create a commonsense policy around this tech's use immediately.
Establish responsible use outside of the workplace.
Don't assume all employees are "AI abstinent." Offering guidance on what information should and shouldn't be included in prompting is paramount to starting a holistic discussion around responsible use. Users should be trained to effectively use the AI tools and interpret their outputs and remain cognizant of AI's limitations and potential risks.
Set boundaries with suppliers.
Whether your team members and suppliers know it or not, any information put into ChatGPT may be accessible to competitors and other third parties. IT teams are already racing to establish rules, but they tend to be internally focused.
As procurement and supply chain teams, it's our job to communicate these expectations to our supplier partners as well. At a minimum, remind suppliers to always implement data anonymization and encryption techniques to protect sensitive information.
Form a tiger team.
Creating effective "prompts" is a key new skill for many, so consider training your teams on prompt creation, or even hiring a "prompt engineer" to optimize results.
You may want to consider implementing procedures to detect and mitigate biases in the AI's data and outputs. This could involve bias audits, fairness metrics and diverse training data sets.
Discuss self-triggering supply disruptions.
What commodity categories are potentially vulnerable to the "panic buying" dynamic? How are other market participants working through this challenge?
These are topics of strategic concern that will matter whether AI is being used by your firm or by other buyers/sellers in the market. Start having these types of conversations with your team now as a preventative measure.
Future Evolution Of Guardrails
As technology advances and business practices change, the risks and legal restrictions surrounding AI will continue to evolve. Consider what future enhancements to AI capabilities may be needed to enable more advanced features like predictive analysis, natural language processing in more languages, bias audits and advanced fraud detection.
There is a lot of AI hype, and it's nothing new for many in procurement and supply chain. Yet, these models are only getting better, and the speed of improvement has never been faster. If we don't have internal discussions and form clear boundaries and specific intentions ahead of this technology's acceleration, we may be stuck with the ideas and recommendations that AI itself generates.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
