How not to fall for AI dupes

If history has taught us anything, there’s nothing quite like widespread buzz to inspire an influx of dupes, and AI dupes are no exception. As far back as the Second Century BCE, growing demand for Greek sculptures among wealthy Romans led to a proliferation of replicas. Today, the dupe phenomenon shows no signs of slowing down, particularly with younger generations — eMarketer published survey data highlighting that 71% of Generation Zers and 67% of millennials say they sometimes or always buy cheaper versions of name-brand products.

The hashtag #dupe, short for duplicate, has wracked billions of views on TikTok as influencers share fashion, beauty and other finds that are “close enough” to the original brands, like Lululemon athletic wear and the viral insulated Stanley Cups. Unfortunately, many buyers are learning the hard way that dupes don’t always live up to their promises.

In fact, Lululemon is among the most targeted brands — its Align leggings, for example, retail for as high as $98, compared to dupes priced in the $20 range. The issue is so pervasive that the brand even held a Dupe Swap at its Century City Mall store in Los Angeles in mid-2023, where customers could trade in dupe leggings for a pair of the authentic product. Lululemon reported that half of the attendees over the two-day event were new customers, with half of them being under 30 years old.

While a disappointing dupe in fashion, beauty, or household goods is low stakes, the same can’t be said for business-critical technologies, like AI. And AI might just be the most buzzworthy technology we’ve seen in a long time.

More airtime attracts more attention

The hype around AI really took off with the 2022 launch of ChatGPT, OpenAI’s viral generative AI tool. AI’s influence seemed to be permeating the zeitgeist. By 2024, a January photo editorial on Wired proclaimed the CES 2024 “The Year AI Ate Vegas.” And tech companies everywhere jumped on the bandwagon, trumpeting AI in their product messaging.

While many companies are genuine in their claims, the exaggerated AI claims –(sometimes called “AI washing” – are equally good at attracting attention. Unfortunately, misrepresenting AI capabilities or giving the impression that AI plays a more significant role than it actually does can also help to establish a competitive advantage by further fueling the attention-getting machine.

FOMO meets FOMU

Investing in AI dupes unwittingly introduces security, financial, compliance and ethical risks to organizations. This leaves many decision-makers in a state of decision paralysis, negatively impacting agility and progress.

While there is a valid fear of missing out when it comes to AI adoption, there’s an equally palpable fear of messing up. Here’s the good news: Armed with the right evaluation criteria, you can protect your company from the fallout of investing in AI dupes.

5 considerations when evaluating AI claims

1. Look for longevity and stability. AI existed long before the current hype cycle. Investigate who the early tech adopters trusted to find the true tech trailblazers.

Evaluate how long “AI companies” have been in business, when AI became a focus, and in what capacity. You should also learn about their financial health to help gauge operational stability and sustainability.

2. Insist on air-tight security. Any AI tech provider you ultimately work with must ensure the security of their AI models and data. They also must have the capability to securely integrate the content you need to fuel your application of their technology.

Industry-standard certifications like ISO 27001 and SOC 2 Type II are non-negotiable. It’s also important to assess compliance with the specific security standards most critical to your organization, such as HIPAA in health care or PCI DSS in financial services.

3. Seek seamless technology integration. You need AI solutions to work flawlessly with your existing tech ecosystem. Your AI tech partners should also support you through integration, with services, clear documentation, and tools that get you running smoothly.

Complexity never scales. Smooth integration with your tech stack from the start will make it easier to grow along with data and processing demands in the future.

4. Understand data strategy and ethics. Garbage in/garbage out should be a real fear with AI data. Your tech partners must be open about what data fuels their models so you can assess quality, accuracy, and relevance.

You’ll also want to understand how they approach data cleaning and validation, what practices are in place to monitor for biases and their strategies for preventing errors and hallucinations.

5. Ask for other opinions. Beyond seeing the technology in action, ask to speak with real customers who have successfully deployed the company’s AI-fuelled tech in their operations. This can be particularly useful in helping you bust through any buzzwords the company uses. For example, if a vendor says their AI is “revolutionary,” ask a real customer what that means in practical terms to their organization. Digging into the specific problems the AI-powered solution is solving for different companies can shed light on performance, reliability, support, and more.

Make AI tech decisions rooted in informed confidence

Anytime you’re weeding through a landscape of noisy claims, it’s important to scrutinize what vendors promise vs what they can actually deliver. AI washing, intentional or not, can significantly damage your organization’s profitability, sustainability, and reputation.

You can protect your brand from the fallout of AI dupes by insisting that your technology partners have the history, talent and, most importantly, proof that their solution is viable for the long term. Due diligence up front ensures you get exactly what you pay for while avoiding the pain of dealing with more than you bargained for.