AI adoption rates appear weak, but current data hides a bigger story

Is AI a bubble?

It's a fundamental yet earthquake question to the minds of many people. But here it is: It's simplifiedtrying to color an unprecedented gray moment in either black or white. And what does the query mean? If you are asking whether valuations of specific AI startups and the companies that offer them are overvalued compared to current finances, there are strong cases to answer in affirmative terms. If the hype around AI asks if they raced ahead of the technical situation when it comes to gaining artificial general information or rapidly destroying the labor market, the answer may be “probably.”

But when we ask if AI will ultimately care and break down history, the science fiction-obsessed Silicon Valley cult's enthralling dreams are found to exist through billions of FOMO fueled venture capital dollars, the answer is undoubtedly “no.”

Anyway, where is AI adoption?

Perhaps the best data that will help you answer any form of “bubble” question is AI adoption rates. Do individuals and businesses actually use AI? Are they profiting? Do they integrate it into what they do every day?

Let's take a look at the business first. Quantumblack, McKinsey's AI ARM, published a report in June showing that 8 in 10 companies use Generative AI.

However, MIT Nanda's report, released two months later, appeared to reaffirm the skeptic's canary call. The researchers analyzed 300 publicly disclosed AI initiatives in various companies, interviewed representatives from 52 organizations, and surveyed 153 senior leaders at major industry conferences.

“Even though enterprise investments in Genai are $3-40 billion, the report reveals an astonishing result of 95% of organizations getting zero returns,” the author reported.

That headline-friendly finding spread quickly, trumpeted by various media sources as evidence that the presumed AI bubble was about to burst. Previously published Quantumblack reports also showed that AI had a similarly unshiny effect on corporate revenues, but at the time it was unable to generate buzz.

Furthermore, since June, companies of all sizes have reduced their use of AI, according to the US Census Bureau.

However, both MIT researchers and Quantumblack analysts will say this is a temporary blip. Given the wide coverage of the media, you may think that they hold a negative view on the practicality of AI in the business world, but the opposition is true. The reality is that both teams are set to gain that AI is already changing the way people work and overhaul the way they operate their businesses.

“For the first time in human history, we can manipulate technology in human languages, not computer science languages,” McKinsey senior partner and Quantumblack's global leader, Alexander Sukharevsky, told Big Think in an interview. “It's always difficult to predict the future,” he warned, adding, “If you look at the current recruitment and drive of an organization to go to change, that's the best I've ever had in my career.”

The two trends will help refuel Sukharevsky and the MIT team's optimism towards AI. First, although most companies have not yet derive concrete benefits from adopting AI, individuals have.

“Shadow Economy” AI at work

“As of late 2024, nearly 40% of the US population ages 18-64 use generated AI,” the trio of economists reported earlier this year. “Twenty-three percent of respondents employed used the generated AI for their work at least once the previous week, and 9% used it every day.”

If these statistics do not appear to be groundbreaking at face value, consider them in a historical context.

“Compared to the launch of the first mass market product of each technology, the work adoption of the generator AI was as fast as a personal computer (PC), and the overall adoption was faster than a PC or the Internet,” the author states. Certainly, PCS and the Internet were difficult to accept first due to the cost and difficulty of setup, but the adoption of AI is consistent with two important products in the current technological age, demonstrating its long-term transformation potential.

The MIT team witnessed this firsthand while investigating the report. The AI ​​initiative was amazed at the business level, but AI itself was widely used among the workforce.

“AI is already transforming jobs rather than through official channels. Our research revealed a thriving “Shadow AI Economy” in which employees use personal ChatGPT accounts, Claude subscriptions, and other consumer tools to automate important parts of their work, often without IT knowledge or approval.

The scale is remarkable. Although only 40% of 40% of the companies surveyed purchased official LLM subscriptions, more than 90% of workers in the companies surveyed report regular use of personal AI tools for work tasks. In fact, almost everyone used LLM in some way for their work. ”

These employees were well aware of what AI needed to succeed in learning and memory, which is the grander enterprise-wide scale. AI needs to be able to retain information over a long period of time and adapt to changing circumstances. This brings us to the second reason that both the MIT and Quantumblack teams are confident of the essential future of AI despite the lukewarm successes so far. AI systems with these capabilities are here and are currently being deployed.

Unleash the agent

The vast majority of AI systems used in today's businesses are easy to implement and shareholder-friendly chatbots. However, these can slightly improve productivity at the worker level – summarizing meetings, creating images, or writing emails will not result in innovative changes to the workflow that drives key returns for your investment.

Chatbots are conversational tools designed to answer questions primarily from a predefined knowledge base. On the other hand, AI agents are sophisticated autonomous systems that allow you to analyze information, make decisions, and take action to achieve your goals. Essentially, chatbots are reactive, but AI agents are aggressive. A chatbot is like a calculator. It's essentially a tool. Agents are collaborators with their own calculators.

“Think about an infinite army of interns who can perform many simple tasks,” Sukharevsky said of Agent AI. “But you need to organize them, you need to upskill them, you need to provide them with the appropriate information.

Agents are in the early stages of implementation, and these could fulfill the large promises of AI.

“Don't expect to see a masterpiece for the first time,” Skalevsky said. “But it actually evolves. What we see today is the worst… If you look at technology a year or two years ago, you see technology today, or technology 12 months later, you get a totally different outcome.”

The rapid progress of agent AI is evidence of a shift towards widespread practical use in businesses. MIT researchers have found more evidence of this trend.

“In the next few quarters, it will be nearly impossible for some companies to lock up vendor relationships and relax,” they write.

The consulting company's sources reflected what researchers were seeing in a trend report published in mid-September. Approximately 55% of clients planned to invest in organisational restructuring over the next 18 months, but it is almost entirely due to AI.

“It's impossible to hide from the influence of AI,” the author writes. “If there is, there is very little if there is no roadmap for AI implementation.”

What does Agent AI look like when integrated into your business? QuantumBlack analysts provided some examples. With eCommerce, agents can view user behavior, shopping cart content, and purchase history to provide product recommendations in real time. In supply chain management, agents synchronized to internal and external data sources can continue to forecast demand, allocate inventory, determine optimal transport, and more, such as determining other things. At retail banks, agents can develop credit risk notes about future clients. This is a task that can currently take two to four days for a human worker. Instead, the person can check the AI ​​job before making a loan decision.

How about us?

If AI agents are essentially digital workers and you can make a massive amount of them, where does it leave human workers?

The broader topic of unemployment in the age of AI is widely covered, heavily debated, and is nearly impossible to predict in the long run. The only thing that can certainly be said is that the nature of the job changes.

MIT researchers have spoken directly from company leaders about the impact of the workforce.

“Organisations that cross the Genai gap are beginning to see the impact of the selective workforce on customer support, software engineering, and management functions. Executives were hesitant to clarify the scope of layoffs for AI, but these companies were 5-20% of customer support operations and management processing work.”

Researchers also sought their opinions from workers in these companies.

“Concerns about the impact of the workforce were far less common than expected. Most users welcomed automation for boring manual tasks, especially as long as the data was safe and results were measurable.”

Sukharevsky briefly explained the role of human workers in the future of agent AI.

“Your role is to find the right team members, the right skills, the right data and oppose the right mission. If you spend enough time, it's [AI] Interns can become your fellow professional one day and add more value. ”

The MIT team feels that the impact of the ongoing AI rollout does not feel equal in all sectors. In healthcare, energy and advanced industries, management did not expect job cuts over the next five years. However, technology and media are currently underway where AI is coding, creating images, creating scripts, creating videos, and creating videos, and creating more in the near future.

But thanks to agents, new jobs and industries will gush out, Sukharevsky said. He's already seen it.

“The newcomers are creating whole new business models that they could never imagine in the past, and they couldn't imagine because the economics of the units couldn't fly.

Sukharevsky emphasized that it will take years for agent AI to be incorporated into a large, established business. Legacy software architecture, human awareness, management, and governance are all barriers to successful recruitment.

“You'll see a variety of results because you need to go 'all in' to get results. It's difficult, expensive, and not immediate. ”

But it definitely comes. QuantumBlack offers pre-built agent and builder software products for companies to create their own creations.

“What we're trying to do is to bridge the gap between AI promises and reality,” Skalevsky said. “This is true because we already have dozens of agent transformations under the belt and are ongoing.”

The answer we seek

Humans generally do not harden for patience. When we face uncertainty, we know the certainty. After the MIT team published the report, many onlookers grasped the concept that AI does not produce financial results. Therefore, the structural changes that AI represents about work and life anxiety may never be realized. But you can read others as you dig deep into the report and reveal that the broad effects of AI are just beginning to uncover.

Change is definitely ongoing. The rise in AI is often compared to the Industrial Revolution. It's someone's guess whether AI will actually result in comparable social upheavals and economic restructuring. The only certainty is that you can't get the answers you're looking for in a single report from the latest recruitment data or over months or years. The Industrial Revolution took place for over 80 years. Even if AI's economic transformation revives four times faster, we still have a long ride.