Amy Webb is the founder of the Future Today Institute, a foresight and strategy firm that helps leaders and their organizations prepare for complex futures. She pioneered a data-driven, technology-led foresight methodology that is now used within hundreds of organizations. Webb is a professor of strategic foresight at the NYU Stern School of Business, where she developed and teaches the MBA course on strategic foresight A regular speaker at SXSW, she is the author of several popular books, including The Big Nine: How the Tech Titans and Their Thinking Machines Could Warp Humanity; The Signals Are Talking: Why Today’s Fringe Is Tomorrow’s Mainstream and The Genesis Machine which explores the futures of synthetic biology. Webb recently spoke to The Innovator about what the future looks like and how corporates do a better job preparing for it.
Q: The Internet as we know it is changing. Generative AI is here. Now what?
AW: Business leaders are caught up in FOMO.[fear of missing out] It might feel like something has developed all of a sudden even though the work has been in progress for quite a while. It becomes urgent and immediate when it reaches a tipping point. This is the case with Open AI’s ChatGPT and other large language models. It seems like every day something new is being announced. Things are moving so quickly it is impossible to make decisions. My message to leaders is it is ok to be skeptical and stop and ask questions. Everyone is talking about it but in my opinion very few know what they are talking about.
The T in GPT stands for Transformer, a type of architecture introduced by Google in 2017 in an academic paper. The idea was for a new architecture for a neural network that was mapped more closely to the brain and could recognize patterns. That is what ChatGPT does: it recognizes what probably comes next in a pattern and if you have a big enough corpus and powerful enough system and some degree of reinforcement with a human in the loop then you can train these systems to do cool things. What is happening right now is that large language models running in the Cloud have enormous computational capacity. This explains the hook up of openAI’s GPT with Microsoft and AWS with Hugging Face and why Google is hoovering up startups.
GPT is one of many models. There are also image generation models like Dall-E. DeepMind built a system to predict the structure of protein folding. There are also multimodal models that can do many things including text, images and manipulating a robot. None of these systems are useful if you don’t understand how they work and what they do. Tools require knowledge and understanding to achieve results. You need to know what to ask the system to do. If you know math, then a calculator can be useful to you but if you don’t know math it won’t. In the year 2023 everyone has a calculator. This tool is so ubiquitous that its invisible. I think that is what is going to happen with generative AI.
Q: Why will this be relevant for business?
AW: Let me give you an example in healthcare. The sector is facing real financial challenges. I took a P&L from 2010 and copy pasted it into text and gave GPT-4 a prompt: find a novel way to reduce overhead by 8% and margins by 15% without reducing head count. Some of what it came up with was generic but here and there it generated answers that many people would be hard pressed to come up with themselves. In the same vein I created a synthetic data set of healthcare data that would normally be shielded by privacy laws – gender, age, first and last name and disease type. The first thing I did was de-identify the patients by giving each a unique five-digit number while keeping the gender and medical information. I did this for the same patients in 2010 and in 2020. I then asked generative AI to find areas where patient outcomes were improving or getting worse and to do an analysis. It did it in 10 seconds. The reason I am bringing this up is the UK government in December 2020 funded a massive very expensive year-long study with the national hospital system and a university involving enormous amounts of data to see if it was possible to write an algorithm to find patterns. I can tell you that it is viable to do it because I did a proof of concept in my living room in minutes. This kind of thing will be true for every industry but in most cases, people aren’t being taught how to take advantage of these tools.
Q: What advice do you have for corporates?
AW: Business leaders are making ridiculously bad decisions right now. They are rushing to partner without really knowing what they are getting into. I talked to one yesterday who said his company was already using the GPT platform. I asked him if the company handed its data to OpenAI. He didn’t know. We have effectively allowed the policing of [large language models] to be done by the companies that develop them, so we don’t know how these systems work or where the sources of data are coming from. What’s more these systems are powerful but confidently incorrect. My advice? Start with a data audit and put the right personnel in place before jumping ahead. Enterprise clients are asking me how to deploy this for their consumers. The better use case is applying the technology to achieve your goals differently. Define your problems. Collect all your data, audit it, secure it, and talk to whoever you are partnering with in the Cloud to get the assurances you are going to need.
Q: What about the metaverse?
AW: The metaverse as [Meta CEO] Mark Zuckerberg envisions it may or may not materialize but using AI to cross over from the physical to digital space and using human data or the human body as the interface is already very much in progress. Head mounted hardware that offers extended reality displays will open a realm of possibilities. Currently if you need a stent in your heart the surgeon uses an ultrasound. Imagine having a digital twin of that heart. It changes the game. It is very likely in advanced economies that surgeons will be wearing an extended reality device to make sure your outcome is better. As digital and physical realities become intertwined, innovation teams in many industries will have an opportunity to pursue new moonshots. If companies develop moonshot ideas and the applications to match it will enable them to positively contribute to the metaverse even as emerging platforms are being built.
Q: The Future Today Institute just came out with a report on 2023 tech trends. Beyond AI and the metaverse what else should be on corporates’ radar?
AW: There is a reason why our 2023 Tech Trends report is named Focus. It is crucial to focus when new signals are forming because some may be lasting and develop into impactful trends, while others might not. Leaders who focus on the trends that matter and adapt to changing circumstances make better decisions and see improved outcomes. Our 2023 Tech Trends report, which is 819 pages and available for free, covers 666 trends and the impact of technologies on 20 different sectors. In each section of the report, we cover what organizations should be doing now to prepare for these trends.
Q: How can companies to do a better job in preparing for the future?
AW: I spoke to a company recently that told me how every five years it initiates a strategic planning process. It brings together the 12 most futuristic people in the company – which is a code for millennials and people interested in tech – and has them meet once a month for 12 months. That is not how you do this. For us it takes six months of regular work with a team on my end of two to three people and a team of five or six people from a company. We don’t just do brainstorming but apply serious rigorous methodology to modeling signals and trends. The process needs to be data driven and then prioritized. We take a quantitative approach and are interested in velocity, impact over time, and how each development relates to other things. That work is intense and foundational and needs to be done before brainstorming scenarios. While storytelling is a part of the process this is not a task for the marketing department. It is about strategy. You need to know where you are today and then do gap analysis and research analysis to backcast to the present and determine what needs to happen and come up with a strategic plan that can be handed off to the strategy team.
This article is content that would normally only be available to subscribers. Sign up for a four-week free trial to see what you have been missing