In recent months, artificial intelligence (AI) has become a favorite buzzword. As the development of AI steadily accelerates, Silicon Valley start-ups and Fortune 500 companies have joined in this industrial revolution. However, excitement, progress and red flags are coming together to reveal a future that is exciting but not necessarily a blessing. During this period, some companies eager to catch the limelight hope to whitewash their status through hype and exaggerate their weak or even non-existent AI technology capabilities.
Compared with non-AI startups, this kind of less bright marketing strategy can help them obtain more substantial seed, Series A and Series B financing. According to data compiled by GlobalData, AI startups raised more than $50 billion in venture capital in the past year alone. In view of the craze caused by technological achievements such as ChatGPT, this number is expected to continue to grow within this year.
With huge amounts of money pouring into startups, the AI hype will only intensify. The U.S. Federal Trade Commission is fully aware of this risk and has warned vendors to be transparent and honest when promoting AI capabilities.
Michael Atleson, a lawyer in the Trade Commission’s advertising practice, wrote in a blog post, “Some products that claim to have AI capabilities may not work as advertised at all. In some cases, the products are even false. The serious harm that may be caused by the loss of functions. Marketers must be aware that, out of concern for the supervision of the Trade Commission, they must not make false or unsubstantiated claims about the efficacy of their products."
In such a complex situation In the real world, it seems increasingly difficult to distinguish legitimate AI solutions from pure marketing gimmicks.
Beena Ammanath, executive director of Deloitte Global AI Institute, said, "Customers must remain appropriately skeptical when faced with claims made by vendors about their own AI products. As with anything, if it sounds good, If it is too unbelievable, then it may be false.”
Donald Welch, CIO of New York University, said that if CIOs and companies fail to see through the propaganda tricks in time, they may face project failure or delay, financial losses, Legal cases, reputational risks or even the outright end of a career. “I’ve seen executives get fired for this, and to some extent they deserved it. ”
Fortunately, there are still several strategies that can help avoid this mistake.
Directly review the vendor’s AI products both It is long and time-consuming, but from another perspective, searching employee information on LinkedIn can help us quickly establish an overall understanding of the supplier.
Ammanath said, "Be sure to carefully check the supplier's employee information. AI experience and education level. Companies developing AI solutions should have such talent, which means they have data scientists and data engineers with extensive experience in fields such as AI, machine learning, and algorithm development. ”
In addition to focusing on employees, CIOs can also look for evidence of the vendor’s collaboration with external AI experts and research institutions. This includes partnerships with universities, participation in industry conferences and events, and support for open source Contributions to AI initiatives, etc.
If the vendor has similar project or application development experience, that is certainly a good sign that it is expected to deliver a high-quality product that also matches the hype.
United States Vira Tkachenko, chief technology and innovation officer of startup MacPaw, a Ukrainian-American, said, “Carefully investigate the historical direction of the supplier. If they are truly an AI expert, they will likely have a track record of research papers in this field or other AI products. ”
To integrate AI technology into products, companies also need to develop well-designed data strategies. After all, data is the “blood” of AI algorithms. . They must be highly dependent on high-quality data. The richer and more relevant the data, the better the AI output will be.
Ammanath pointed out, "AI systems must be driven by large amounts of data, so companies should definitely have A good data strategy explains how much data they collect and where it comes from. ”
Another focus is whether these companies have made sufficient efforts to comply with regulatory requirements and ensure high data privacy and security standards. With the General Data Protection Regulation (GDPR) and With the introduction of advanced data privacy regulations under the California Consumer Privacy Act (CCPA), organizations must be transparent about their data practices and provide individuals with control over their data. Failure to do so may require liability for vendors. A big question mark is placed on my ability.
Although the slogan can be extremely tempting, the customer must still seek verification in a gentle manner. Ammanath believes that "asking the right questions and asking the other party to provide evidence to support the conclusion of the promotion can be said to be critical to deciphering marketing claims and determining whether the product is truly powered by AI."
In evaluating so-called AI-driven When developing a product or service, CIOs can ask how the model was trained, what algorithms were used, and how the AI system adapts to new data.
Tkachenko said, "Everyone should ask the supplier what libraries or AI models they use. Everything they show may be based on simple OpenAI API calls."
Management and Matthias Roeser, partner and global head of technology at technology consultancy BearingPoint, agrees. He added that CIOs should thoroughly understand the components and framework of a product or service, including assessing "ethics, bias, feasibility, intellectual property and sustainability."
By asking, CIOs can learn more about the product’s true capabilities and limitations, and decide whether it’s worth spending real money on.
Startup companies are at the forefront of innovation. Many of them have indeed broken through the boundaries of AI possibilities through their own efforts, but many of them are just blindly exaggerating their capabilities. Want to make quick money.
Vlad Pranskevicius, the Ukrainian co-founder and CTO of Claid.ai, a startup company under Let's Enhance, said frankly, "As the CTO of a machine learning company, I often encounter cases of AI hype, especially in the startup field. Even more." He also noticed that the recent situation has become more and more exaggerated. After all, in the gold rush caused by AI, many people want to take advantage of the hype cycle to make a fortune.
But Pranskevicius believes that as regulations on AI become more stringent, AI hype will be controlled in the near future.
It’s not uncommon for customers to buy AI solutions that don’t live up to their name, and it’s likely not that the CIO did anything wrong. Welch said that this may be "the consequence of poor corporate leadership syndrome. Business departments lose their minds in the face of crazy marketing, and the IT team repeatedly blocks it to no avail, and is ultimately forced to clean up the mess."
In order to prevent this from happening In such a situation, a collaborative culture must be developed within the organization. In this culture, the opinions of technical professionals should be valued and they should be allowed to elaborate on their opinions and evidence.
At the same time, CIOs and technical teams should also build reputations within the company to ensure that their opinions are included in the decision-making process. To achieve this goal, they should strive to demonstrate their expertise, professionalism, and soft skills.
Max Kovtun, chief innovation officer of Sigma Software Group, said, "I don't see anything wrong with CIOs checking the hype of AI. The biggest problems often come from business stakeholders or the founders themselves being dismissed by what appears to be innovative and cutting-edge. The publicity has gone to the head, and we are desperate to pursue AI. So the most important question is how to avoid becoming a victim of AI hype in the context of the raging fire."
When comparing products and services, they must be evaluated with an open mind and a comprehensive look at their essential attributes.
Tkachenko believes, “If the only advantage of a product or service for you is AI, you should think carefully before purchasing. For example, it is best to study its value proposition and functional features. Make sure you understand the benefits beyond AI before making a decision."
Welch also agreed, "Would I buy a product because it was written in C, C++, or Java? Of course not, I What I want to know is whether the supplier can maintain this batch of code well and whether it can survive in the cruel market competition for a long time."
A thorough evaluation can help organizations determine whether the products or services they plan to purchase meet the requirements of goals, whether they can provide the expected results.
Kovtun emphasized, “The more complex the technology, the more difficult it is for non-professionals to understand, and it is even impossible to verify whether the technology should be used and whether it is meaningful. Therefore, before deciding to introduce AI technology into the business, it is best to It is best to hire experts with rich experience and knowledge in the field of AI. Otherwise, your efforts may not bring the expected benefits."
Learn about AI-related products and The latest information also helps CIOs make informed decisions. In this way, they can identify loopholes in each other's propaganda and follow up on new ideas and technical achievements in a timely manner.
Art Thompson, CIO of the City of Detroit, believes, “I feel that the current level of AI education is not enough.”
He suggested that CIOs should do their homework to avoid falling into the technical trap of promising levels that exceed actual delivery capabilities. Once this happens, "the time wasted in rebidding and changing products will make it difficult for employees to keep up with the changes, and everyone's enthusiasm for investing time in learning new technologies will also be greatly reduced."
In addition, learn about the latest AI Trends can also help CIOs anticipate regulatory changes and industry standards, stay one step ahead of compliance and maintain a competitive advantage.
Of course, keeping abreast of the latest information is not just the responsibility of the CIO alone. BearingPoint’s Roeser noted that “you should also educate your team or hire technical experts to add relevant capabilities to the human mix.”
New regulations can simplify CIOs The process by which we determine whether a product or service truly uses AI technology. The White House recently released an AI Bill of Rights, which includes guidelines for how to responsibly design AI systems. In the next few years, more relevant regulations may be introduced.
Ammanath pointed out, "The purpose of these actions is to protect consumer rights and even all mankind from the potential harm of technology. We need to predict the potential negative impacts of technology and ensure that risks are reduced."
Companies are often willing to discuss new technologies themselves and emphasize the potential benefits, but often downplay the negative impacts.
Philip Di Salvo, a postdoctoral researcher at the University of St. Gallen in Switzerland, said, “When a technology creates a craze, we tend to ignore the harmful effects it may have on society. Research shows that companies are driving AI-related discussions, and Among them, technological determinism still dominates.”
This view of technology as the main driving force behind social and cultural change is likely to ignore useful discussions of moral and political implications, and instead supports More marketing-oriented arguments. As Di Salov puts it, this creates "a fog of public opinion that gives technical solutions and even their producers greater room for maneuver and opportunities to evade responsibility."
To solve this problem, he believes that it must be The public realizes what AI is not and cannot do.
Di Salvo concluded, “Most of the AI applications we see today – including ChatGPT – are basically built around large-scale statistics and data analysis applications. This sounds like a boring definition, but it helps people avoid over-associating with the "intelligent" expression here. We need to pay attention to the real problems in the field of AI, including issues such as bias and differential treatment, and not be intoxicated in hypothetical and speculative long-term visions. ”
The above is the detailed content of Here's one way to rewrite it: The rise of AI is unstoppable, here are nine ways to help you avoid overhyping it. For more information, please follow other related articles on the PHP Chinese website!
![[Web front-end] Node.js quick start](https://img.php.cn/upload/course/000/000/067/662b5d34ba7c0227.png)
