In recent months, artificial intelligence has been everyone’s favorite buzzword. Both Silicon Valley startups and Fortune 500 companies believe AI will steadily pick up pace to revolutionize their industries. But excitement, progress, and red flags like AI washing are developing just as well. Some companies are desperately trying to get on the gravy train, but want to capitalize on the hype, so despite the fact that the AI they’re actually employing is minimal or non-existent, Exaggerating AI capabilities.
This dubious marketing strategy helps them receive larger seed, A, and B funding rounds compared to non-AI startups. Last year alone, AI startups raised more than $50 billion in venture capital funding from him, according to GlobalData, and that number is expected to rise this year given the frenzy surrounding ChatGPT and others.
Given the money poured into these startups, the AI-washing phenomenon will only intensify. The U.S. Federal Trade Commission is fully aware of this danger and has warned vendors to be transparent and honest when advertising AI capabilities.
In a blog post, Michael Atleson, an attorney for the FTC’s advertising practice, said: “In some cases, this lack of efficacy may exist regardless of what other harm the product may cause. You should know that false or unsubstantiated claims about product efficacy are our bread and butter.”
In this complex landscape, it can be difficult to distinguish between legitimate AI solutions and marketing gimmicks.
Veena Amanas, Executive Director of the Deloitte Global AI Institute, said: “Like anything, if the story sounds too good, it’s very likely.”
Donald Welch, CIO of New York University, says that failure of CIOs and their companies to find the right answers can result in project failures and delays, financial losses, lawsuits, reputational risk, and ultimately layoffs. He said he could face it. “I’ve seen executives fired, and I can’t say it was the wrong decision.”
Luckily, there are some strategies you can use to avoid making mistakes.
AI-powered businesses need skilled employees
Screening companies that claim to be using AI can be a long and time-consuming process. However, something as simple as performing a LinkedIn search can yield valuable insight into your organization’s profile.
“Look at the level of AI experience and education that vendor employees have,” says Ammanath. “Companies developing AI solutions must have the talent to do so, which means they have data scientists and data engineers who have deep experience in AI, machine learning, algorithm development, etc. ”
CIOs can not only survey employees, but also look for evidence of collaboration with external AI experts and research institutions. This category includes university partnerships, participation in industry conferences and events, and contributions to open source AI initiatives.
It’s also a good sign if the vendor has experience with similar projects and applications, as it indicates that they can deliver quality results.
“Check your supplier’s history carefully,” says Vira Tkachenko, chief technology and innovation officer at Ukrainian-American startup MacPaw. “If the company is an AI expert, they will most likely have a history of research papers in this area or other of his AI products.”
Look for a well-crafted data strategy
Companies that truly integrate AI into their products also need a well-thought-out data strategy, as AI algorithms require it. They need to work with high-quality data, and the more generous and relevant that data is, the better the results.
“AI systems are underpinned by so much data that these companies need to have a well-structured data strategy and be able to account for the amount and source of data collected,” said Ammanath. increase.
Another thing to check is whether these companies are doing enough to comply with regulatory requirements and maintain high data privacy and security standards. With the rise of data privacy regulations such as the General Data Protection Regulation (EU GDPR) and the California Consumer Privacy Act (CCPA), organizations need to be transparent about their data practices and give individuals control over their personal data. there is. If this does not happen, it is a red flag.
demand evidence to support a claim
Buzzwords may be glamorous, but they serve to quietly seek evidence. “Ask the right questions and demand proof of your product claims, which is crucial to peeling the language off marketing and sales to determine if a product is truly AI-powered,” he said. says Ammanath.
CIOs evaluating specific products or services that they believe use AI ask how models were trained, what algorithms were used, and how AI systems adapt to new data can do.
“You should ask vendors what libraries and AI models they use,” says Tkachenko. “It could all be built with simple OpenAI API calls.”
Matthias Roeser, global technology leader and partner at management and technology consulting firm BearingPoint, agrees. He adds that components and frameworks need to be fully understood, and that evaluations should include “ethics, bias, feasibility, intellectual property, and sustainability.”
This inquiry will help the CIO learn more about the product’s true capabilities and limitations, thereby helping the CIO decide whether to purchase the product.
Focus on startups
Startups are at the forefront of innovation. But while many of them are pushing the boundaries of what is possible in the field of AI, others are exaggerating their capabilities to get attention and money.
“As the CTO of a machine learning company, I often come across cases of AI washing, especially in the startup community,” says Vlad Pranskevičius, co-founder and CTO of Ukrainian-American startup Claid.ai. However, he has noticed that the situation has become more serious lately, adding that the phenomenon is especially dangerous during hype cycles like the one we are experiencing now, as AI is perceived as the new gold rush. I was.
However, Pranskevičius believes AI washing will be curbed in the near future as regulations around AI tighten.
Build a reputation as a technical expert
It’s not uncommon for companies to acquire questionable AI solutions. In such situations, he is not necessarily at fault with the CIO. It could be “a sign of poor leadership in the company,” he says, Welch. “The business fails the marketing hype and dismisses the IT team, who needs to pick up the rest.”
To prevent this from happening, organizations need to foster a collaborative culture in which technical experts’ opinions are respected and their arguments are exhaustively enumerated.
At the same time, CIOs and technical teams need to build a reputation within the company so that their voices are more likely to be incorporated into the decision-making process. This requires demonstrating expertise, professionalism and soft skills.
Max Kovtun, Chief Innovation Officer at Sigma Software Group, said: “A bigger issue may be that business stakeholders and entrepreneurs are putting pressure to use AI in all its forms because they want it to look innovative and cutting edge. So the right question is how to avoid being an AI washer under the pressure of entrepreneurship.”
Beyond buzzwords
When comparing products and services, it is imperative to take a thorough look at their attributes and evaluate them with an open mind.
“If AI is the only advantage of your product or service, you should think carefully before subscribing,” says Tkachenko. “We encourage you to study its value proposition and capabilities, and start working together only when you understand the benefits of the program beyond AI.”
Welch agrees: he asks. “You might want to understand that as part of your due diligence on whether they can maintain the code, viability of the company, etc.”
A thorough evaluation helps an organization determine whether the product or service it plans to purchase is fit for purpose and likely to provide expected results.
“The more complex the technology, the harder it is for non-experts to understand it to the extent that they can verify that the application of that technology is correct and makes sense,” Kovtun says. “If you decide to utilize AI technology for your company, it is advisable to hire knowledgeable professionals with experience in the AI domain. It may not work.”
Up-to-date information on AI-related products and the issues surrounding them can also help CIOs make informed decisions. In this way, they can identify potential mistakes they may make and take advantage of new ideas and technologies at the same time.
Art Thompson, CIO of the City of Detroit, said:
He recommends that CIOs do their due diligence to avoid falling into the trap of new or experimental technologies that promise more than they can deliver. In that case, “the staff won’t be able to accept the change because it takes time to rebid and replace the product,” he says. “Not to mention the difficulty of spending time learning new techniques.”
Additionally, being informed on the latest AI-related matters helps CIOs anticipate regulatory changes and emerging industry standards, helping them stay compliant and stay competitive.
And CIOs aren’t the only ones who need to stay up to date. “Educate your team or hire an expert to add relevant functionality to your portfolio,” says Roeser of BearingPoint.
Additional regulatory measures for AI
New regulations underway could simplify the task of CIOs trying to determine whether products and services employ real-world AI technology. The White House recently issued an AI Bill of Rights containing guidelines for responsibly designing his AI systems. Also, more regulations are likely to be issued in the coming years.
“The premise behind these actions is to protect consumer rights and humans from potential harm from technology,” says Ammanath. “We need to anticipate the potential negative impacts of technology to mitigate risk.”
Ethics should not be an afterthought
Companies tend to influence the discourse of new technologies, often emphasizing potential benefits while downplaying potential negative consequences.
Philip Di Salvo, Postdoctoral Fellow at the University of St. Gallen (Switzerland), said: “Surveys show companies are driving the AI debate, with the tech deterministic debate still prevailing.”
This belief that technology is the main driver behind social and cultural change can obscure discussions about ethical and political implications in favor of more marketing-oriented arguments. As Salvo puts it, this creates a “fog of controversy in the form of further obscuring and de-accountability of these technologies and their producers.”
To address this, he says, there is a key challenge of telling the public what AI can and cannot actually do.
“Most AI applications we see today, including ChatGPT, are fundamentally built around large-scale statistical and data analytics applications,” says Di Salvo. “This may sound like a boring definition, but it helps avoid misunderstandings of what “intelligent” refers to in the definition of “artificial intelligence.” We need to focus on real issues such as prejudice, social classification, and other issues rather than hypothetical, speculative long-term scenarios. “
