Generative AI: The beginning of (another) technological revolution
We are likely witnessing the birth of a technology as significant as the internet itself. The growing demand for generative AI will have ripple effects across industries, with winners and losers emerging over time.
KEY TAKEAWAYS
The expected growth in AI suggests investment opportunities for companies that facilitate its infrastructure, such as hardware manufacturers and cloud computing vendors
AI will drive enormous productivity gains and incremental economic growth as it is adopted
As a new industry, there are risks related to regulation, data privacy and cybersecurity
The internet, Wall Street and just about every manager in every business have been buzzing about AI chatbots since a wave of new launches began in November 2022. It took just five days for a popular AI chatbot to reach 1 million users, compared to 10 months for another popular social media company and 3.5 years for a common streaming service (Figure 1). Tech giants and startups alike are racing to develop ever more intelligent AI algorithms to get a toehold in the next generation of computing. In short, we are likely witnessing the birth of a technology as significant as the internet.1
What is generative AI?
These new AI-driven chatbots use technology known as generative AI, which is machine learning taken to an entirely new level. When we talk to Alexa or Siri, we are actively engaging with machine learning. Machine learning is also what drives weather prediction and statistical modeling across industries. It drives the estimates of likely outcomes we see for everything from betting odds in sports to the probability of an earthquake in a vulnerable city.
What makes generative AI models such a breakthrough is their ability to discern context and create new content, drawing from massive amounts of data. When starting a sentence, generative AI can predict with high accuracy how it will end. To build the models, some of the world’s best data scientists and computer engineers trained models using 570GB of books, webtext and articles from across the internet.
Whereas machine learning could identify pictures of animals or patterns from data, generative AI can generate images of animals based on all available similar images or write unique paragraphs about historical events with appropriate contextual references to time, place and participants. Ultimately, generative AI may change how a broad range of jobs are performed across almost every industry.
Researchers using generative AI can quickly summarize complex topics, a development that worries educators that want to see students produce original work. AI will help humans with increasingly intensive tasks and generate new content or ideas. The quality of the models can produce outputs that are, in certain cases, indistinguishable from human-generated content.
AI will likely require that we revisit the concept of “original content” given its ability to “invent”. In medicine, AI may be able to predict what sequence a “next new medicine” might take by looking at prior molecules or compounds.
Figure 1. Worldwide Google Searches for ‘Artificial Intelligence’
Source: Google as of April 18, 2023.
Potential opportunities and risks
PwC estimates that AI could contribute up to $15.7 trillion to the global economy by 2030. A large chunk of this is estimated to come from productivity gains ($6.6 trillion). Examples include AI helping educators grade exams, software developers to write or debug code, job seekers to write resumes, or online content providers to sell and advertise products and services.
Generative AI will be a key building block for many other future technologies. Generative AI is disruptive too, capable of driving enormous productivity gains and incremental economic growth in the years ahead. Winners and losers will emerge from across all sectors of the economy. And that includes a huge impact on workers.
In his congressional testimony in May, Sam Altman, the CEO of OpenAI, the company that created ChatGPT, suggested the rapid regulation of generative AI — remarkable given that few tech CEOs seek regulatory oversight. He suggested the formation of a new government agency charged with licensing large AI models and empowering it to revoke that license for companies whose models don’t comply with government standards. He also proposed safety standards for AI models, independent audits by experts and evaluating and extinguishing dangerous capabilities like acting independently.
Indeed, regulation, patents, copyright and data privacy concerns are key risks to watch. Intellectual property is another major issue. Copyright laws vary worldwide, but if someone uses generative AI to write a paper, will the sources of the information be correctly attributed? Another concern is generative AI creating false and even dangerous conclusions, drawing from inaccurate data. People do this too, but generative AI can create and spread false information faster. The technology’s code-writing ability arms cyber criminals with an easier way to write malicious code or create new phishing scams. Then there are social implications, especially when it becomes difficult to discern between human and machine interactions. As this revolutionary technology rolls out, these issues will require humans to create policies, practices, checks and balances.
As the technology develops, there will be winners and losers across industries. In the race to establish dominance in AI, larger, well-established companies have more to lose when rolling out generative AI because of reputation risk. Rushing out products or services that are misleading or low quality would be a hit to their reputations. But the benefits will be bountiful for those who get it right and can use AI to support their businesses. There will be billions spent on learning, introducing, embedding and assessing the impacts of generative AI.
Figure 2. Valuations and EPS Growth Rates for AI, Robotics and Technology Shares
Source: Factset as of April 18, 2023. Indices corresponding to each column AI: Indxx Artificial Intelligence and Big Data Index; Robotics: ROBO Global Robotics and Automation Index. The indices are unmanaged. An investor cannot invest directly in an index. They are shown for illustrative purposes only and do not represent the performance of any specific investment. Index returns do not include any expenses, fees or sales charges, which would lower performance. Past performance is no guarantee of future results. Real returns may vary.
What to do now
Pure AI equity investment opportunities are few and far between. The most advanced AI development platforms are either privately held or housed within large companies with diverse sources of revenue. That said, we see potential growth opportunities in companies that provide the computing and infrastructure that power AI (Figure 2). These include hardware manufacturers making advanced semiconductors and infrastructure providers such as cloud computing vendors.
Secondary parts of the generative AI ecosystem stand to benefit as well. Semiconductor capital equipment companies that produce the machines that manufacture AI-optimized chips should see increased demand (Figure 3). Cybersecurity companies are another obvious initial beneficiary that may be underrated.
And finally, we will be highly mindful of the companies that adopt the technology in their products and operations. Increasingly sophisticated AI models will make robots more productive and autonomous vehicles more reliable. Services-oriented companies that can embed AI effectively into day-to-day processes will begin to improve their enterprises’ quality, value, efficiency and profitability.
We expect, over time, to see trends emerge and to see companies report on how they are using generative AI in their operations. Intuitively, AI will be able to track the use of AI. That’s the whole point. This technology will build upon itself. Firms that do not embed this technology into their business practices may see their competitive positions will erode.
Figure 3. Shares in Industry Groups Tied to AI in 2023-to-date
Source: Bloomberg as of May 31, 2023. Indices are unmanaged. An investor cannot invest directly in an index. They are shown for illustrative purposes only and do not represent the performance of any specific investment. Index returns do not include any expenses, fees or sales charges, which would lower performance. Past performance is no guarantee of future results. Real results may vary.
Download your copy
Discover our Mid-Year Wealth Outlook 2023 insights and themes.
