When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
Since the dot-com bubble of the early 2000s, it’s hard to argue that any other tech innovation has grabbed the headlines more than generative AI (GenAI).
The stage was set in late 2022 when OpenAI threw their generative AI chatbot, ChatGPT into the air, allowing the world to witness what this technology was capable of.
However, the euphoria at the time was so high that many pushed the innovation aside, calling it all hype. Businesses were skeptical about having their names inked side-by-side with GenAI tools.
But the story has changed. Generative AI is now the go-to phrase for about every business that wants to drive traction or be perceived as innovative and progressive (even if it is more a marketing tactic than a technical one).
But away from the rhetorics of marketing and PR grandstanding, generative AI may well become the backbone of business value derivation.
Apart from reports pointing towards its increasing adoption, new evidence suggests that GenAI could help businesses generate more value from their investments.
In this piece, Techopedia takes a closer look at latest studies and speaks with experts to understand how generative AI moved from hype to value in less than two years.
Key Takeaways
- Generative AI has rapidly transitioned from being dismissed as hype to delivering significant cost savings and revenue gains for businesses that have adopted it.
- Human resources and supply chain management are major areas benefiting from generative AI, but inaccuracies in outputs for these domains can lead to legal issues and revenue costs.
- Despite the surge in generative AI adoption, most companies are still unprepared to implement it responsibly.
- Businesses building proprietary models from scratch are currently seeing the highest value.
Generative AI Now Drives Revenue Gains
According to a recent McKinsey Global Survey on AI, enterprises are already reaping significant benefits from adopting GenAI in 2024. The survey findings reveal that Human Resources is one aspect of business where GenAI use guarantees the largest share of cost reduction.
On the revenue front, respondents cited meaningful revenue increases of over 5% in supply chain and inventory management by leveraging GenAI capabilities.
Early adopters who have heavily invested in GenAI are reaping the biggest rewards, attributing over 10% of their organization’s Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) to their use of GenAI technologies.
Other industry reports corroborate these findings. Boston Consulting Group’s survey revealed that about half of the companies surveyed experienced over 10% in cost reduction just by deploying gen AI.
Commenting on the findings, Julien Salinas, CEO at AI privacy by design platform, NLP Cloud, told Techopedia that this is just the beginning of the gains businesses will derive from gen AI.
“I think this is only the beginning as many businesses still don’t clearly understand what powerful capabilities generative AI is bringing.
“However, it’s important to note that the current price businesses are paying for GenAI is not the “real” price they should pay for it.
“Most AI vendors like OpenAI are selling their AI models at a loss. But it won’t last forever.”
Roman Kucera, CTO and Head of AI at Ataccama, told Techopedia:
“Businesses that use AI have the potential to gain significant competitive advantage. This will be the case for the foreseeable future, and companies not using AI risk becoming obsolete.”
However, he cautioned, “With everyone ultimately using AI, it will not provide huge returns of investment for all of them — some will be able to leverage it better, some worse. Finding the right fit, the best AI use cases to provide the most value will be critical.”
HR Suffers Most from GenAI Inaccuracies
According to McKinsey, as enterprises reap the rewards of generative AI, they also grapple with the risks associated with data privacy, bias bias, and intellectual property infringement, to model management issues such as inaccurate outputs and lack of explainability. Additionally, security vulnerabilities and potential misuse pose another key challenge.
One of the well-known issues with GenAI is that it prioritizes plausibility over accuracy and can create inaccurate outputs — a problem that has been well-documented and known as AI hallucination.
The risk of AI hallucination mostly affects the human resources side of businesses and has resulted in revenue costs in many ways, McKinsey found. This is mainly because when GenAI output related to HR processes like hiring, payroll, or compliance is inaccurate, it could lead to legal issues, fines, and reputational damage for the company.
Just like HR, the supply chain is not safe from this risk. Businesses are wary of using GenAI for inventory management as any inaccuracies in demand forecasting, inventory optimization, or logistics planning could severely disrupt operations and impact the bottom line.
Issues around AI hallucinations result from feeding AI with obsolete, conflicting, or mislabeled data, Kucera said.
“Training AI with the wrong data can lead to inaccurate responses, compounding existing data issues. Before leveraging AI, companies must ensure they have high-quality, curated data. Failing to do so risks amplifying errors rather than gaining insights.”
Beyond wrong training data, Arti Raman, CEO and Founder of Portal26, cites malicious data poisoning and biased training data as another reason why businesses struggle with inaccuracies in their AI response.
“Even a small amount of flawed data can significantly skew outcomes in AI chatbots, with disastrous consequences in critical sectors,” Raman said. However, she highlighted a possible solution:
“Projects like Google’s Knowledge Vault demonstrate a method for validating information by cross-referencing multiple sources, evaluating credibility, and only incorporating data meeting stringent accuracy thresholds into model development.”
Investing in such solutions is crucial to avoiding inaccuracies, rather than sidestepping GenAI altogether, she noted.
Most Companies Aren’t Prepared to Address AI Security Concerns
Despite concerns about security risks surrounding AI, there is still a remarkable surge in GenAI adoption.
Most large companies see generative AI as crucial, yet lack preparedness for responsible implementation, reveals a McKinsey flash survey of over 100 firms with over $50M in revenue. 63% prioritize generative AI adoption, but 91% feel unprepared to do so ethically, exposing a significant readiness gap.
Security leaders also have significant worries – 77% say regulatory uncertainty around GenAI impacts their deployment decisions according to KPMG.
There are concerns around data privacy violations, copyright infringements from training data, and the potential for GenAI to enable large-scale misinformation.
Businesses that Build LLM from Scratch Make the Most Gain
There are three distinct strategies organizations adopt when deploying generative AI capabilities, according to McKinsey:
Takers
These are organizations that primarily use open-source GenAI models and tools from big tech companies and startups like Meta’s Llama 3 and xAI’s Groq-1. Around 40% of the respondents are categorized under this category. These takers leverage pre-trained models for tasks like content generation, language translation, and data analysis without much customization.
Shapers
About 35% of companies fall into the category. They are companies that take existing large language models and fine-tune or retrain them on their proprietary data to create custom-GenAI solutions tailored to their specific needs and use cases.
Makers
Lastly, around 25% are investing heavily in developing their own GenAI models from scratch, training large language models on proprietary data using substantial computational resources. These tend to be tech giants and large enterprises with the expertise and infrastructure to build cutting-edge GenAI capabilities in-house.
The survey reveals that makers see the highest value from GenAI currently, with over 15% of their EBIT attributed to this technology versus 10% for takers and shapers. However, the takers benefit from faster time-to-value while shapers balance customization with reduced development costs.
The Bottom Line
Generative AI is no longer mere hype – businesses are beginning to experience its tangible impact, and there’s no indication they’ll relinquish this powerful technology anytime soon. However, a valid concern arises: Will the pursuit of profit overshadow the resolve to address the security, safety, and ethical issues woven throughout gen AI?
Only time will reveal the answer, but before deciding whether to adopt pre-existing GenAI tools, customize them, or create proprietary models, organizations must consider the data or intellectual property they wish to protect. While bad actors may find it more challenging to infiltrate proprietary large language models (LLMs), even this approach doesn’t guarantee foolproof data security.
In the race to harness GenAI’s potential, businesses must strike a delicate balance between innovation and responsible implementation, ensuring that ethical considerations remain at the forefront of their decision-making processes.
