ChatGPT is turning two. Here’s a look at the state of the generative AI union.
OpenAI’s ChatGPT is celebrating its second year of general availability on Nov. 30, which means it’s time to take stock of generative AI’s progress in 2024. And what a whirlwind it’s been.
Two years ago, few people knew GenAI was possible but today roughly 40% of Americans have used the technology. To put that adoption trajectory into perspective, that growth doubles the 20% of Americans who used the Internet within two years of its launch, according to the National Bureau of Economic Research.
You read that right: The Friggin’ Internet.
The pace of GenAI innovation has been unprecedented. In 2024 alone, OpenAI broke new ground in LLM reasoning, while Meta introduced the first open frontier class model. Google meanwhile conjured a breakthrough in GenAI-fueled podcasting.
And Anthropic launched tools that help users create and modify content in a separate window, as well as the ability for computers to use computers. (You need to see it to understand it).
AI Agents Shine Bright
Excitement over AI agents is palpable, as organizations seek to amplify not just employees’ productivity but operational efficiency. At a high level, AI agents are pieces of software code that execute tasks to achieve a preset goal. Most AI agents can “think,” or reason, plan and learn from feedback.
However, agents also take many forms. First, AI agents can include digital assistants that help consumers. Think software bots that can book travel and handle other transactions, etc. Then there are enterprise agents, which can work individually or as part of teams (multi-agent architectures) to automate whole workflows or entire business processes. Eventually, these agents will be ablet to “self-heal,” identifying errors and correcting course.
And while it’s premature to argue that AI agents will automate an entire business, organizations are certainly interested in their potential. Eighty-two percent of leaders surveyed by Capgemini said they expect to use agents to automate the generation of emails, software code and data analysis.
Small Language Models Can Do Big Things
Some people are wary of the money hyperscalers are pumping into GenAI infrastructure, software and talent. However, understanding the motivation is critical; these companies are investing in super intelligent systems—a big leap beyond the everyday content creation applications most organizations are pursuing.
The reality is organizations needn’t spend millions of dollars building or licensing large language models (LLMs). Rather, small language models (SLMs) running in hybrid IT environments provide more than enough AI firepower to satisfy most targeted business use cases.
“You’re going to see a set of use cases that emerge where a small, less accurate model will be a lot better than what you had and probably good enough,” said Mindy Cancila, vice president of corporate strategy at Dell Technologies, in a recent webinar.
Moreover, SLMs’ smaller footprint means they can run on anything from servers to laptops to smartphones, fed by data housed anywhere from corporate datacenters to public cloud services and out to the edge, where breakthroughs in model compression and performance will enable high-quality inferencing at low latency.
Progress Will Bring Productivity Windfalls
Plenty of research suggest that GenAI has boosted productivity across organizations. In truth, actual results are difficult to quantify, according to academic Ethan Mollick, an expert on GenAI adoption within organizations, who noted that business leaders report little AI use and few productivity gains outside of niche use cases.
Mollick further argues that organizations must perform research and development into understanding organizational AI use gauge productivity and other progress metrics. And those R&D analyses have yet to be codified—even by the consultancies who get paid to do the work.
“Nobody has special information about how to best use AI at your company, or a playbook for how to integrate it into your organization,” Mollick said.
Still, consultancies continue to find positive metrics from GenAI investments and adoption.
For instance, Ernst & Young LLP found that senior leaders whose current budgets for AI investments are 5% or more of their total budgets saw higher rates of positive return across several key areas compared with those who spend less than 5%.
Those that allocated 5% or more of their budgets outperformed their lower spending peers 76% to 62% for employee productivity, 71% to 55% for product innovation and 73% to 47% for creating competitive advantages, according to EY.
This suggests that organizations can ill afford not to increase investments in GenAI. Their competitors certainly will.
Of course, balancing business strategy with IT investments to get the desired business outcomes is never easy. But you don’t have to go about it alone; trusted advisors like Dell are here to help.
Learn more about the Dell AI Factory.