AI Trends In 2022: What’s Real And What’s Hype? Hear From The Experts
The end of the year is a time not just for predictions of top trends but also to watch for the biggest hype and most misleading recommendations that get dished out to business leaders. There’s no scarcity of these in the artificial intelligence (AI) space.
I asked several industry leaders five actionable questions for 2022. Inspired by Tribe of Mentors, the bestselling book by Tim Ferris, I gave the traditional questions a slight twist. (“If you want uncommon clarity and results, ask uncommonly clear questions,” advises Ferris.)
I sent the following questions to data and analytics experts, asking them to respond to up to three of them:
- What's a rarely discussed data and analytics trend that will materialize in 2022?
- What's the biggest hype in AI that will fizzle out in the new year?
- What is one bad recommendation you often hear being given to business leaders?
- What's one pandemic-triggered change you will continue in your organization in 2022?
- What is one data/analytics book you've gifted the most to non-technology leaders?
This article is organized around these questions. Feel free to skip around to those that interest you the most.
1. What's a rarely discussed data and analytics trend that will materialize in 2022?
A strong digital product sense can be the next big thing. This hyper-personalization will be aided by a dramatic increase in efficient data preparation. Ethics in AI could also get a much-needed boost.
- Having a strong digital product sense will be the next big thing as more businesses move online. Digital-first business models rely on personalization to improve customer experience and retention. Product scientists who can quantify user experience on digital products, design online experiments, and use machine learning on clickstream data will be in high demand. – Nikhil Sikka, Associate Director, Advanced Analytics and Insights of Wayfair
- We’ll see the rise of systems for efficiently cleaning and structuring data. Today, armies of individual data scientists at large companies create their own notebooks and manually work on datasets with no shared infrastructure or coordination across teams. We’ll see the rise of tools that will standardize and dramatically accelerate the creation of these data-munging pipelines. – Charles Fisher, CEO of Unlearn.AI
- Conversations around ethical AI and ethical data analytics will attract more attention from academia and industry. Hopefully, this emphasis will work its way toward policy-makers and politicians. The recent press coverage on whistleblower documents about giant social media companies could help accelerate meaningful change in 2022. – Hadi Hosseini, Assistant Professor at the College of Information Sciences and Technology, Pennsylvania State University
2. What's the biggest hype in AI that will fizzle out in the new year?
Experts in AI often consider the term artificial intelligence to be overused and misinterpreted. For example, data leaders in pharma point to how AI in drug discovery is blown out of proportion by the media and why it should be interpreted in context.
- Our current artificial intelligence is not really intelligent. Most of the work today involves people running huge datasets to create complex statistical systems. This is very valuable, but the intelligence is ours. We are grouping a lot of non-intelligence work into the term AI, and the title of data scientist is becoming too generalized. This creates a false narrative and underestimates the work needed to materialize the ‘real’ AI. – Elwin Loomis, Head of Digital of Bremer Bank
- While AlphaFold2 from DeepMind is an amazing achievement and a great solution to a 50-year-old challenge, the media has overhyped it. Biology and drug discovery are extremely complex with many open questions, and this solution plays a small but critical role in this broad space. This reminds me of the media hype around the original human genome project. I hope the hype fizzles so that we can appreciate it for what it is and move on to solve the next big problem. – Brandon Allgood, Chief AI Officer of Valo Health
- AI for drug discovery won’t be a singular category, given how democratized it already is. Companies that offer it must be judged on the quality of their clinical candidates in addition to their machine learning or novel data platform. Machine Learning (ML) in drug discovery is part of the scientific tool kit and will continue to evolve with academic and for-profit partnerships. – Milind Kamkolkar, Former CDO of Cellarity
3. What is one bad recommendation you often hear being given to business leaders in the data and analytics space?
We’ve heard cliched claims such as “data is the new oil” or ‘AI is a strategic enabler.” Some of these are marketing devices that have outlived their usefulness, while others are well-intended advice that can still cause harm.
- Data and analytics leaders are often told to build their organizations to “support” business functions. This is a reactive, outdated mindset. AI and analytics should inform a business strategy from the ground up. Analytics leaders should level up from “order takers” to “thought partners”—for example, using analytics to track the evolution of trends as they unfold, learn from those insights, and make strategic business decisions based on real-time data. – Nikhil Sikka, Associate Director, Advanced Analytics and Insights of Wayfair
- Businesses are often advised to focus on strategic use cases. While this is sound guidance for getting an AI program up and running, it becomes limiting as initiatives proliferate. Organizations struggle to align and synchronize across their many “strategic” initiatives. This leads to technical debt, redundant solutions, and disjointed customer experiences. Organizations must rapidly pivot from localized point solutions with AI to orchestrated and scalable capabilities across the enterprise. – Todd James, Chief Data & Technology Officer of 84.51˚
- The entire ethos captured by the phrase “data is the new oil” is terrible. It has led to the practice of hoarding lots of data with the hope that “insights” will be discovered via some computer magic. Data can be extremely useful for answering questions—but you need to know what questions you want to answer. Better advice is to first identify your questions, then go out and collect the data you need to answer them. – Charles Fisher, CEO of Unlearn.AI
4. What's one change triggered by the pandemic that you look forward to continuing in your organization in 2022?
Covid-19 brought many unpleasant changes, but it triggered some shifts we may have never discovered otherwise. From an increase in flexible work policies to a rise in compassion for team members’ unique personal circumstances, there are things the pandemic forced us to prioritize.
- We’ve seen an increase in compassion and recognition for coworkers' individual circumstances. Working at home alongside other family members leads to many clashes, and some of this is on video (e.g., when your CEO's child joins an investor call and they don't bat an eyelash). The pandemic has really opened peoples' eyes to the hidden struggles and joys everyone has. – Brandon Allgood, Ph.D., Chief AI Officer of Valo Health
- We have created two fully remote working time blocks globally over the summer and winter breaks. It creates a great change of scene and enables colleagues to spend time in other places or countries with family members they might not have seen because of the pandemic. It is a hugely popular policy. We never would have thought that possible before the pandemic. – David Benigson, CEO of Signal AI
- Flexible work hours enable people from various socio-economic or personal backgrounds to be able to work on their terms. This is key for advocating diversity by allowing individuals—for instance, mothers with young children or individuals from underrepresented groups—to advance in their careers. – Hadi Hosseini, Assistant Professor at the College of Information Sciences and Technology, Pennsylvania State University
5. What is one data/analytics book you've gifted the most to non-technology leaders?
Book recommendations could be many, but which were the ones the leaders purchased for their friends? Here are two suggestions that cover AI applications and their societal impact.
- Framers: Human Advantage in an Age of Technology and Turmoil by Kenneth Cukier et al. Imagination can’t be programmed, and this book sets out an interesting thesis on what it means to be human in the age of AI. – David Benigson, CEO of Signal AI
- Reprogramming the American Dream by Kevin Scott of Microsoft provides clear insights into the business impact of AI applications, but it goes further to discuss societal implications and how they may be addressed. – Todd James, Chief Data & Technology Officer of 84.51˚
AI’s deepening influence on humanity
As AI evolves, its influence on humanity continues to rise. People often focus on AI’s ability to automate and amplify tasks but underestimate its more profound impact on society.
“Very few human creations have had the kind of impact as AI,” says Loomis. He compares it with the invention of language—a “tool” that has changed the trajectory of humans and helped birth civilizations. Today, we are still taking baby steps with AI. However, unlike early humans, we are waking up to the fact that AI is not just a tool but will weave deeper into our society.
“I hope 2022 will be the start of this realization, where we don’t just create new technical practices for AI but also understand how it shapes us. This should alert us to the fact that this is the time to lay the guardrails—the checks and balances needed to guide this change into something greater and not dystopian,” concludes Loomis.
Note: The quotes shared in this article are opinions of individuals and may not necessarily reflect the position of their respective organizations.
Post a Comment