4 principles for retail’s use of AI

What retailers need to know about implementing artificial intelligence
VP, Retail Technology & Cybersecurity; Executive Director, Center for Digital Risk & Innovation

Artificial intelligence applications and technologies are reshaping nearly every aspect of retailers’ business operations and customer engagement. Consumer goods take a long journey within a retail company on the way to purchase, and AI is fundamentally changing that journey.

Take a shirt, for example. Its design might be influenced by AI-enabled trendspotting on social media, or supported by AI-based image generators. The retailer might utilize AI in its work with suppliers to optimize product sourcing and ensure compliance with sourcing restrictions. The various modes of transportation for the shirt’s journey from supplier to retailer might be influenced by AI-enabled decision-making.

AI Working Group

Join our AI Working Group to learn more about AI issues and the development of practices and guidelines for the use of AI within retail.

After it arrives in a retailer’s warehouse or distribution center, the shirt might be sorted or moved using AI-enabled automation technologies. Retailers might use artificial intelligence to support marketing efforts, both in-store and online, or perhaps to facilitate personalization or customization of it. Developing effective promotions or discounts for the shirt might be informed by AI technologies.

After purchase, the retailer’s customer service and return activities are likely to be supported by AI. And all throughout this journey, retailers might be using AI tools to prevent theft and fraud, and protect customers’ information against cyber threats.

Necessary precautions

Looking at these widely disparate current and potential use cases together, the net result is a fundamental transformation of the retail business model and ongoing changes to the customer experience.

Many retailers are moving quickly to integrate AI into businesses — according to market research firm IDC, the retail sector ranks second among all industry sectors globally in its spending on AI technologies. Companies that are not integrating AI into their strategies and business operations risk being left behind by their competitors and by new market entrants.

Retailers must also proactively address a set of key risks when they are looking to use artificial intelligence. AI applications might produce results that are erroneous or that lead to biased outcomes. A large proportion of the general public is still distrustful of AI-informed decision-making and recommendations.

In addition, retailers and other companies face numerous risks from adversarial use of AI, including novel cyber threats and deepfake-enabled disinformation campaigns. And retailers run risks of running afoul of current or future policies and regulations, not just within the United States but in other countries where they may operate.

Dive deeper

Check out NRF's Center for Digital Risk & Innovation to learn more about artificial intelligence in the retail industry.

Principles for use

Given this general context, it is critical for retailers to be proactive in their internal governance of AI and ensure they are using these technologies in ways that support their core values, mission statements and business objectives.

In support of these efforts across the sector, the NRF Center for Digital Risk & Innovation recently published a framework report, “Principles for the Use of Artificial Intelligence in the Sector,” the first articulation of core AI principles for the retail sector. These principles, which are anchored on the trust-based relationships that retailers have with customers, fall into four key categories:

  1. Governance and Risk Management: Retailers should develop strong internal governance of AI tools and capabilities as a foundational basis for managing risks and ensuring that AI delivers expected benefits.
  1. Customer Engagement and Trust: Retailers should be transparent about their uses of AI that have a legal or similarly significant effect on a customer, establish safeguards to prevent unlawful discrimination against protected classes of individuals, and align their governance of consumer-facing AI applications with existing internal privacy, cybersecurity and other data governance policies.
  1. Workforce Applications and Use: Retailers should engage in ongoing oversight and review of AI applications that could directly impact employees or that are used by the workforce to support business needs.
  1. Business Partner Accountability: Retailers should establish clear guidelines and expectations for business partners that are providing AI tools, data sets and services.

These principles — discussed in greater detail in the report — can be used by retailers of all sizes to inform their AI strategies and governance. They might be particularly relevant for smaller and medium-sized companies that could be earlier in their journey of using AI applications and technologies.

They are also the starting point for work that NRF intends to do in the coming months and years, in close coordination with retail leaders in NRF’s AI Working Group and in partnership with other AI stakeholders in the public and private sectors.

For additional information on the AI Principles and opportunities to get more engaged with NRF’s work on AI and other emerging technologies, please reach out to the NRF Center for Digital Risk & Innovation at cdri@nrf.com.

Related content

Center for Digital Risk & Innovation
 
NRF Center for Digital Risk and Innovation
NRF's hub around key technology issues that have significant policy and risk management implications for retailers.
Read more
AI is set to innovate last-mile delivery for retailers in 2025
 
Delivery truck
Startups are using cutting-edge technology to enhance delivery speed, reduce costs and offer greater flexibility.
Read more
How AI can solve the discoverability problem for retailers
 
Woman working at a desktop computer.
Addressing the gap between the language of the customer and the language of brands.
Read more