AI's double-edged sword
(This article is part of a Series on Artificial Intelligence (for board members and senior executives).
Artificial Intelligence (AI) is a double-edged sword that presents as many dangers as opportunities. AI can transform business, revolutionize processes, increase efficiency, and drive innovation. Organizations actively adopt AI for customer service, sales and marketing, predictive analytics and more. AI can also present new, unexpected challenges.
The potential threats posed by AI are very real and can disrupt operations and create chaos that is harmful and costly. However, the biggest risk is the risk of disruption if you don't actively incorporate AI into your business.
Let's start with a closer look at each of these risks.
Potential threats from AI
New Cyber Security Threats
AI gives hackers new sophisticated tools, especially with new open source AI technology. We see more cyber criminals, especially foreign bad actors, using AI to create highly effective phishing campaigns and cyber attacks. Since OpenAI released ChatGPT in November 2022, phishing emails have increased by 1,265%.
The biggest challenge is deep fax. Using AI, hackers can impersonate anyone, including CEOs and decision makers. Take the case of a Hong Kong finance worker who was duped into paying $25 million by deepfakes. The worker was suspicious when he received an email, purportedly from the CFO, asking for the funds to be released, but the clincher came when the worker joined a Zoom call in which the CFO and his Showed to his colleagues. The operator was the only human on the Zoom call. All the rest were AI-generated deepfakes.
We are entering an AI arms race where we need AI-powered solutions to protect against hackers using AI. Make sure your security vendors are using the latest AI technology and that your management team understands the potential threat posed by AI.
AI Threats to Corporate Reputation
If not properly implemented and monitored, AI can put corporate reputation at risk.
Take the case of Air Canada, which damaged its reputation for the way it handled AI-powered customer service failures. A passenger traveling to attend a family funeral failed to request a bereavement fare, so when he went back to Air Canada to request a discounted payment, an AI-powered chatbot told the passenger, “No problem. ” However, there was a problem, as Air Canada's policies clearly state that the discount must be requested prior to travel. Air Canada chose to fight the case in court rather than provide the small discount requested. Fighting the case in court damaged Air Canada's reputation.
Be careful where and how you deploy AI tools and be diligent in monitoring their performance. More importantly, consider the impact on your reputation when handling AI-related issues.
AI can present legal risks.
Like any technology, AI is imperfect, and those flaws can lead to legal problems..
For example, HelloDigit, a financial technology company, didn't guarantee customers money savings and overdrafts. A flawed AI algorithm results in the company pocketing a portion of the interest and overdraft fees and penalties for customers. This led the Consumer Financial Protection Bureau (CFPB) to take action against Hello Digit. Hello Digit will likely have to defend its AI solution in court.
AI enables some wonderful things, but just because AI can do it doesn't make it legal. For example, Facebook's (now Meta) AI facial recognition technology can identify people in a photo with you. What sounds completely innocent — here you are in a photo your friend took — actually violates Illinois' Biometric Privacy Act. The settlement cost Facebook $650 million.
AI can pose operational risks.
Automating operational processes using AI also requires consideration of management oversight and human behavior.
In 2021, Zillow's stock price plummeted primarily because they relied on a flawed AI algorithm to predict house prices. Zillow was buying homes based on these predictions. AI can be a powerful tool for predictive analytics, but even AI can only make predictions based on historical information. Events like the COVID-19 pandemic certainly disrupt predictive analytics and thus hurt Zillow.
Any operational change must also consider the human factor. People have to adapt to new AI processes. For example, UPS developed ORION, an AI-powered platform for routing packages and trucks. But for Orion to work, UPS had to convince workers how to operate according to the new system. At first, the employees did not like the adaptation. It took time for employees to get used to the new AI-powered approach. Ultimately, Orion reduced fuel and maintenance costs and enabled drivers to make more deliveries. However, ORION almost failed because the employees didn't really want it.
Risk of disruption
The biggest risk is not aggressively adopting AI when you have competitors.
According to IBM, 42% of enterprise-scale organizations have deployed AI, and 40% are exploring AI adoption models, including 59% that have accelerated their AI investments.
AI is already dramatically impacting sales and marketing, product development, service operations and supply chain management. Companies that fail to leverage AI to create a competitive advantage risk being left behind (see “AI's Competitive Advantage”).
AI opportunities provide competitive advantage.
To use AI to best competitive advantage, organizations must determine where artificial intelligence can deliver the greatest returns (see “CEO's Guide to Generative AI: Proactive Advice for 2024”).
McKinsey research shows that creative AI use cases typically fall into four areas: customer relations, sales and marketing, software engineering, and research and development.
Some of the most promising areas of opportunity presented by AI to organizations are:
- Sales and Marketing – AI is improving marketing and increasing sales by increasing customer engagement. For example, Capgemini uses an AI platform to track buyer intent through the sales funnel, increasing the number of qualified leads by five times. Starbucks uses AI to analyze customer data to customize offers and applies dynamic pricing to increase customer loyalty and average order value.
- Product development – AI is invaluable in predicting customer desires and behavior, making it ideal for product development. StitchFix uses AI to curate personalized clothing for consumers, analyzing preferences, fashion trends, feedback and other factors to customize outfits and influence inventory decisions. .
- Service Operations – AI can facilitate customer service. For example, Unilever's AI system reduced customer service response times by 90 percent. Companies also use AI for predictive analytics for applications like scheduled maintenance. GE Aviation, for example, reduced unplanned downtime by 50% and saved $1 billion in maintenance costs.
- Supply chain management – As the UPS example shows, AI can improve shipping and logistics. Maersk uses AI for real-time shipment tracking, making it easy to alert customers in case of delays. Deli's Food Service in Brazil uses AI to optimize dynamic pricing, increasing their gross margin by 1.1% (a huge amount in a low-margin business) by managing changing market prices.
Any organization that wants to compete in today's market will embrace AI, but C-level executives should keep their eyes open. While AI has risks that need to be managed, AI can bring tremendous benefits, giving any company a competitive edge over larger rivals.
If you care about how AI determines the winners and losers in business, how you can leverage AI to benefit your organization, and how you can manage AI risk, Contact me on this number. glenngow.com. I write and speak about how senior executives, board members, and other business leaders can effectively use AI. You can click to read past articles and get notified about new articles. “Follow” button here.