We seem to be trapped in a loop of hype and hysteria around generative AI, its various acronyms that confound non-technologists (GPTs, LLMs, GANs, etc.), and their implications for the economy, the future of work, and, it seems, humanity itself.
When I talk to fellow CEOs, though, they are most concerned about two things: First, how do we ethically, safely, and securely implement these new technologies? Second, how do they translate them into applications that drive value for their company?
Generative AI has become the center of a public discussion around its ability to create original content (text, images, video, code, etc.), going beyond the analytical and predictive capabilities we have come to expect from AI. With their easy-to-use interfaces, chatbots powered by generative AI (like ChatGPT and DALL-E) blur the line between human and computer, automating technical and often-tedious tasks and dramatically increasing productivity.
It is easy to understand why they have captured the popular imagination.
And not just the imagination, of course. According to Pitchbook, annual VC investments in generative AI have increased by 425% since 2020 to $2.1bn, with $11 billion invested in the first quarter of this year alone. As the media commentary has exhaustively made clear, the potential of these technologies—for both good and ill—is tremendous.
As I read these futurist musings, though, I come back to those original two questions.
First and foremost, ethical and regulatory issues must be our top consideration. Concerns by top scientists and developers of artificial intelligence around the pace of its deployment should give us all pause. New risks from cyberattacks, disinformation, fraud, and unfettered surveillance are proliferating because of these new technologies.
These new risks highlight the importance of security as a concern. Many generative AI tools (like ChatGPT) are not enterprise ready. My friend Julia de Boinville at Scale AI points out that “most organizations are not going to find out-of-box solutions that meet their needs.” Data security concerns, the need for customized fine-tuning on your proprietary data, and the tendency for these models to produce false or misleading results will mean that most organizations will need customized approaches that address these concerns.
We must move carefully and deliberately to ensure we are taking all necessary precautions.
Second, working within those security parameters, what are the use cases in which we can begin to develop and deploy these new technologies? AI, is of course, already helping us run our businesses more efficiently and productively. Advances in machine learning algorithms, the proliferation and connectivity of data, the availability of cloud-based AI services, and the increasing processing power of computers have made these advances possible.
In a classic case of necessity being the mother of invention, the pandemic and its disruptions to both supply and demand dramatically accelerated adoption of these technologies. Systems that provided greater visibility, data analytics and the predictive power of AI became essential for those who could not otherwise forecast the next week, much less the next month or quarter.
But where do we go from here?
Supply chain management is one area that seems ripe for investment. Today, companies like Amazon, UPS, and DHL routinely use AI-powered systems to optimize supply chains, improve delivery times, and reduce costs. At AlixPartners, we have developed a Global Trade Optimizer TM tool that can take a company’s entire supply chain (at all levels), identify nodes with critical exposures, and apply predictive modeling and monitoring of leading indicators for identified risks. Given the disruptions in recent years to supply chains, these capabilities have never been more critical.
Customer service is another area in which many companies (and their customers) have had terrible experiences. As a result, many companies are reluctant to replace call centers with generative AI-driven chat interface. But co-pilot features, which empower call center operators with LLMs to answer questions and increase productivity are proving extremely promising. Beyond call centers, HSBC is using a similar AI tool to enable more efficient service and optimize pricing.
And generative AI is revolutionizing customer relationship management databases, unlocking hitherto siloed information. These technologies can help reveal where knowledge, skillsets, and experience lie within an organization, making it quickly accessible and functional. They can also more easily map client relationships and networks of influence, which has the potential to revolutionize sales and marketing.
AI can in fact be more than a figment of the IT department. It can be a valued member of your workforce, whether you are serving customers, sorting supply chains or a million tasks in between. As my colleagues Angela Zutavern and Ted Bililies wrote recently, the real power of technology generally, and AI specifically, is not as a replacement for human labor, but as our collaborator. The world keeps getting more complex, dynamic, and harder to predict. Navigating this environment means adapting quickly and making real-time decisions with the data available (which is always growing). Applying AI tools to help make these decisions is the only practical approach. However, simply making these surprising connections and insights is not enough. Applying our human wisdom and judgment will remain critical.
With all the hype around generative AI, artificial intelligence has had its Netscape moment. Just as the Netscape browser allowed the average person to interact with and see the potential of the internet, chat tools show us just how far AI has come and let us glimpse their tremendous potential.
The challenge for us all is to deploy these new technologies ethically and securely and, of course, to identify the right use cases in which to invest.