Avoiding AI Washing: Key Insights for CMOs

AI washing is on the rise, posing a challenge for CMOs to identify genuine AI solutions. We provide actionable strategies and real-world examples to help you navigate this situation effectively.

Understanding AI Washing and Its Impact

Not all that glimmers is gold. The term “AI washing” describes companies exaggerating their AI capabilities. According to Securities and Exchange Commission chair Gary Gensler, AI washing includes “false claims to investors by those purporting to use those new technologies.”

The financial sector has faced significant scrutiny. Delphia (USA) Inc. and Global Predictions Inc. were fined a total of $400,000 by the SEC for marketing AI-enabled investment predictions that did not exist.

This issue extends beyond finance, with FactSet reporting that 179 S&P 500 companies mentioned AI in their earnings calls in the first quarter of 2024, far surpassing the five-year average of 73 companies.

However, merely mentioning AI does not equate to authentic AI-driven operations. As Microsoft’s venture arm M12 managing partner Michael Stewart notes, “putting AI into a slide deck is easy, but it doesn’t translate into sustainable competitive advantages.”

Identifying Genuine AI Solutions

CMOs should request detailed explanations and case studies from potential AI partners. For example, OpenAI successfully leveraged AI for customer segmentation and targeting, resulting in a 20% increase in campaign ROI.

Such detailed accounts of AI implementation can provide valuable insights into the authenticity and effectiveness of the proposed solution.

Historical Usage and Transparency

Examining the timeline of a company’s AI adoption is crucial. Toby Coulthard from Phrasee emphasizes that companies transparent about their AI usage before the popularity of models like ChatGPT are generally more credible.

For instance, IBM has documented its AI journey since 2016, showcasing a clear evolution and continuous improvement in its AI technologies.

Verification of Data Sources

Evaluating the data sources an AI company uses is vital. Michael Stewart’s four D’s framework—data, dividends, distribution, and delight—emphasizes the importance of access to essential data and its impact on AI solutions. Salesforce, leveraging proprietary data sources for predictive analytics, improved customer satisfaction and retention rates significantly.

Continuous Learning and Adaptability

AI solutions should be designed for continuous learning and adaptability. For example, chatbots used by Zendesk evolve with new data and outperform static models over time, ensuring long-term success.

Red Flags Indicating AI Exaggeration

Vague Terminology

Companies should clearly define the type of AI they employ. Ambiguous terms can indicate AI washing. If a company claims to use AI for “enhanced performance” without providing technical details or results, it’s a red flag.

Dependence on Third-party Models

Firms that rely solely on third-party AI models without customization may not offer sustainable solutions. Timothy Bates advises verifying the extent of customization and performance metrics. An instance where a company uses third-party models with minimal customization can be indicative of superficial AI claims.

Inconsistent or Inflated Claims

Exaggerated claims without independent validation are another indicator. A startup promising 95% accuracy in predictive analytics should substantiate this with customer testimonials or independent verification. Consistency between promises and actual deliverables is key.

Sudden AI Adoption

Companies that suddenly shift to AI-centric narratives during periods of high AI market hype might be overstating their competencies. Analyzing the historical context and the long-term strategy behind such claims is essential.

Ensuring Long-term Viability of AI Solutions

Sustainability Assessments

Assess the sustainability of AI models considering their computational costs, energy consumption, and environmental impact. Sustainable practices, like those of Google which uses energy-efficient AI models, contribute to the long-term viability of AI solutions.

Alignment with Business Goals

Ensure that AI solutions align with strategic business objectives. Walmart integrated AI into its supply chain management, leading to significant cost savings and improved inventory management. Such alignment with broader business goals ensures the AI solution’s relevance and efficacy.

Regular Evaluation and Updates

Establish a process for ongoing evaluation and updates. Regularly reviewing performance metrics and making necessary adjustments ensures that AI solutions remain effective amid changing market dynamics.

Conclusion

Navigating the complexities of AI adoption requires vigilance and thorough evaluation to avoid AI washing and ensure the long-term viability of AI solutions.

It may be tempting to engage in AI washing for short-term wins, but ultimately, the truth will prevail, potentially causing significant reputational damage and undermining the long-term success of your business.

References

• Curry, R. (2024, May 10). AI washing: A Microsoft VC’s warning about dubious, and rising, corporate artificial intelligence claims. CNBC.
• AI Ethics Lab. (2023). Identifying Fake AI Solutions.
• Gartner. (2023). Ethical AI and Sustainability.
• GlobalData. (2023). Integrating AI into Business Processes.
• McKinsey & Company. (2022). Evaluating AI’s Long-term Viability.
• Smith, J. (2023). The Risks of Sudden AI Adoption. Harvard Business Review.

—Inspired by: AI washing: A Microsoft VC’s warning about dubious, and rising, corporate artificial intelligence claims, CNBC, Rachel Curry