The AI Snapshot 2022
Navigating Emerging Technology in Modern Business
Preface
Our assimilation with artificial intelligence becomes more interwoven and complicated every year. In 2018’s “​​The Anatomy of an AI System,” AI was described as “simultaneously a consumer, resource, worker, and a product”. As a truly emerging technology, artificial intelligence (and, more imperatively, machine learning) challenges us all to think outside our traditional thinking patterns — and into a new age of blockchain, speech recognition, and deep learning.
​
Inspired by the many reports circulating as AI’s emerging technologies rapidly evolve, such as Stanford’s AI100, The AI Snapshot 2022 focuses on AI's applications and implications for businesses, investors, and AI startups. AI Partnerships Corp. collaborated with our network of over 120 AI Affiliates to develop this report, bringing together thought leaders from data analytics & extraction, to healthcare & manufacturing, who are changing the AI landscape as we know it today.
​
In celebration of the past century of AI, we have prepared this report, The AI Snapshot 2022. Sources include Google searches, academic reports, and the 120+ AI companies in the AIP Affiliate Network — the people in the forefront of the industry in 2022 and beyond.
Why is AI worth discussion
The goal of AI development is to support human lives. We see this in the hopes of self-driving cars and automated business analytics. It is prevalent in the news surrounding the commercialization of Big Data and global climate strikes against large tech companies.
​
These developments lead to fears of how far AI will go, creating conversations surrounding job security to robot domination, most found in online communities, such as Reddit, Slack, and Facebook.
​
Online forums provide a unique combination of widespread attraction and accessibility by creating content highways between niche and extensive communities. However, without knowledge of what AI is and how to use it, honest discussion about its future cannot happen.
​
To us, artificial intelligence is the future of our business and personal worlds. It strives to make our lives better, from healthcare to finance to marketing.
“AI will only get better and better over time, crossing into thresholds where its utility becomes more realistic. Good AI is AI that you don't notice or see; it works in the background without the user knowing there's AI there at all.”
— Don Bosco, Founder @BHuman.ai
AI Innovations
In the past five years alone, AI advanced from potentially harming the environment around us to significantly reducing carbon emissions. By improving computational efficiency — a known cause of AI’s carbon footprint — AI scientists and researchers continue to develop new pathways for neural networks.
​
The focus on the global climate challenge is just one of many innovations in the past decade. From AI co-producing a mainstream album, "I AM AI." by Taryn Southern to AI art-making generating half a million USD of sales in 2018, the innovations and conversations surrounding AI have come far beyond our imagination.
With vast quantities of satellite, aerial, and drone-based data about the Earth being collected daily, AI is a critical tool for rapid classification and interpretation of remotely sensed data and for reducing required quantities of data storage and transmission across the Earth Observation industry.
Earth observation AI is a rapidly expanding domain that includes application of AI technology throughout the planning, collection, and processing of data that benefits commercial organizations and governments around the world driving both innovation and contributing to sustainable growth and resource management.
“While there will be rapid advances in data labeling and machine learning, the most powerful advances in the field of AI will be the mainstreaming of AI, which will occur through the design, development, and implementation of custom AI solutions. Staying competitive in AI requires speed, agility, and daily surveillance for advances in tools, resources, and strategies.”
— Michael Geisen, Director of Sales @Cloud Brigade
Innovations in 2021
As popularity in AI adoption spreads, it’s exciting to see which companies blaze new paths forward for technologies like computer vision and natural language processing (NLP). These innovations result in improved AI education, creating more intelligent AI overall.
​
The following examples take a look at which AI innovations we paid attention to during 2021 — and the implications it leaves us with for 2022.
Coding & Programming
AI has always been a prevalent innovator in programming. It is no surprise that AI Coding continues to propel innovation in NLP for tech companies large and small. Just as NLP can analyze human language, Github’s CoPilot can analyze and “read” code using a special variant of a NLP model, OpenAI’s autoregressive language model, GPT-3.
​
Innovations in AI-powered chips also continue to progress in industries such as cybersecurity, introducing features such as one-time DRAM scrambling and I/O virtualization. Ambarella leads the industry with their newest chip, CV52S, sporting their new portfolio based on the CVflow architecture.
​“Federated learning is a promising & innovative solution to empower ML algorithms by taking advantage of external computing machines without sharing data.”
—Yuqi Li, Founder @AIpaca
Digital & Virtual Humans
Chatbots have been a part of smart marketing for years. They are so widespread that anybody with a Facebook Business Page can set up a no-code Messenger bot. However, Facebook took it a step further in 2021 with BlenderBot 2, a state-of-the-art open-source chatbot that maps text inputs on-demand while encoding/decoding its’ dialog history, internet searches, and long-term memory.
​
As AI becomes more human-like than ever, the concern of human mimicking, such as deepfakes, also becomes widely discussed. Innovations in virtual humans emphasize this concern, including releases like MetaHuman Creator, despite the engine being focused on studio production.
​“Human language is an extremely difficult scientific and industrial domain. In the nearest future, I expect that we will finally see the models that include non-linguistic components together with the linguistic ones.”
— Sasha Uritsky, CEO @nmodes
Ethical AI
We have still yet to address the ethics of progressive AI in a serious, intelligent manner. Each issue, including bias, fairness, safety, transparency, and accountability, has stirred up new nonprofit organizations like the Royal Society and the Open Philanthropy Project to take the reins in addressing the ethics of progressive AI.
​
While ethical AI as a topic has many touchpoints to be aware of, the significant tangential consequences include uneducated legislation, data privacy breaches, and discrimination in protected practices like hiring.
​
These severe repercussions inspired several companies to come together in 2016 to form the Partnership on AI, an organization dedicated to advancing positive outcomes for people and society.
“At Sama, ethical AI has been at the core of our business model, and it remains an essential component when we make any business decision. This means that we are not only committed to providing ethical training data but also focused on finding ways to improve the AI supply chain through employee wellness programs, access to living wages, and training resources"
— Alex Shee, VP Corp Dev & Strategy @Sama
Implications for 2022
Despite the apprehension around Ethical AI and job security, the future is far from bleak for AI. US job markets predict a strong recovery in 2022, thanks, in part, to AI.
In a few years, AI will be in all products and services, impacting every business and role. Responsibility will continue to be the most considered value for end-users, creating a hyperfocus need for companies to prove their expertise using AI responsibly.
​
Luckily the beauty of AI is how much data processing it can do in so little time. As AI adoption becomes the norm for businesses worldwide, ethical AI becomes less of a threat to us all. The growing accessibility of small and wide data means no more monopolization — and thus furthered democratization — of data.
​
That same AI can also create multi experience software, or software that combines data and generates new content from it.
“AI helps sales teams make significantly more informed decisions, which eventually shows in every company's top line. It can reduce the resources involved to gain insights, reveal patterns that were otherwise hidden, and take timely action informed by more varied and larger data on an ongoing basis”
— Malay Upadhyay, Chief Customer Experience Officer @SalesChoice
The future of AI is both inevitable and necessary. Ultimately, using AI will be critical to scaling at pace with the technological world for businesses across any industry in 2022. Just like remote work became the norm in 2020, AI will soon become the norm of our daily lives.​
Hurdles Standing in AI's Way
Outside of the prejudices we have developed for AI through media and movies, many other hurdles stand in our collective way for widespread AI understanding and adoption. Unfortunately, the specifics of what holds us back are nothing new to the world of tech.​
Hurdle #1: Businesses Aren't Ready
Many companies are excited by AI, yet they are uneducated on how the technology works. This is not an issue for most tech startups, but AI startups are not software companies. AI is not a "set it and forget it" type of tech.
​
AI has incredible power in the world of customization for any business, but even the best AI takes time to learn the nuances of the company. In that way, AI is not much different from hiring a human employee, other than the pace at which it obtains data and the speed at which it improves.
​
The waiting period that businesses experience with AI onboarding is not only a time-consuming process; it is expensive, too. When the high costs of operations finally show through the veil of excitement, some companies back out of their initial AI investments before the work has even begun.
​
When Google's DeepMind initially released, its impressive learning pace matched that of humans'. With how deep learning models work, though, that pace of learning will continue to increase exponentially — as will the cost of technology development.
​
Soon, DeepMind will far surpass a human learning pace, accelerating data collection and processing of all kinds, but not without the steady stream of capital from Alphabet.
“Most business models aren’t ready to adopt AI. Adopting AI is a capital-intensive R&D process. Many companies seemed to believe the hype on AI "stealing" jobs, but it couldn't be further from the truth; there will always be demand for autonomous solutions.”
— Craig Ganssle, Founder & CEO @Cadre-AI
Hurdle #2: Poor Data Management and Infrastructure
Unfortunately, most businesses do not know how to set up their data infrastructure clearly and concisely. The gap between leadership and data scientists is so large that more than 50% of all data scientists were looking for new roles in 2020.
​
Moreover, because businesses cannot link and leverage their data, they cannot understand it. Businesses using AI are 2.3x more likely to be high performers, yet only 29% of businesses are prioritizing AI initiatives linked to business goals.
​
In order to adopt AI to create better insights from data, data needs to be in one place as much as possible. Central repositories like data lakes are becoming more and more popular, allowing data to be more readily available for AI solutions, among other benefits such as scalability.
"Data lakehouses collate the company data that stays isolated in databases, user events, and documents. Building bridges across the data offers new insights that are hidden in plain sight."
— Ozan Eren Bilgen, CEO @Base64.ai
Hurdle #3: Investors Aren't Educated
In the past two years alone, AI companies have attracted almost $100 billion in funding from venture capitalists. C3.ai IPO'd at 100%+ gains the day it launched, while AI-generated NFTs continue to grow in popularity.
Even with these impressive numbers, investors fear AI-related risks. According to venture capitalist Peng T. Ong, managing partner at Monk's Hill Ventures, investors don't see the value in AI startups. His beliefs — shared by other VCs worldwide — start with the basics, like how AI startups are at the highest risk of failure and expand past the hype of AI.
​
Seasoned VCs will not stare in awe of new robots or tech. They know that, in time, AI will become a constant in our daily lives and that only successful startups will withstand the tests of time.
Hurdle #4: AI Businesses Aren't People-First
AI businesses face unique challenges in the marketplace. They are not a software company, they are not "just" a tech company, and roughly 60% of “AI startups” do not actually use AI in their businesses. It is more complicated than ever to break into new audiences with sentiments from investors like those listed above.
For the AI startups that do use AI, ROI is hard to showcase without historical proof of concepts. To compete in today's startup ecosystem, AI startups do not just need great tech; they also need wider audiences to reach through marketing.
However, with Big Data companies like Facebook on the hot seat in 2021, data privacy is a forefront concern for most. Due to this, AI businesses face an uphill battle of reeducation and reintroduction to new technologies. AI, when ethically produced, is safe — but companies must stay on top of retraining their AI to stay that way.​
“Data is the biggest challenge for businesses, and any innovation (or regulation) that impacts the data journey remains our focus. It’s imperative that businesses become ‘AI-native.’ It needs to be their core competency to succeed over the next decade.”
— Monish Gandhi, CEO & Founder @Gradient Ascent
AI Adoption By Vertical
Unsurprisingly, AI adoption skyrocketed due to COVID-19, and we were still seeing the effects in 2022. While 52% of companies began their AI adoption plans in 2020, 67% also pushed their goals into 2021. As time goes on, businesses will undoubtedly continue to begin or advance their plans to adopt AI; such is the nature of new tech adoption.
​
Healthcare
​
22% of AI companies are experiencing a surge in demand due to healthcare's constant need for supportive infrastructure and efficient practices. AI companies like MarkiTech and AI Vali are creating the future of telehealth with remote patients and physician speech recognition, respectively, while other startups support healthcare staff to automate shift scheduling.
Digital Analytics & Security
Conversely, to the worries of data privacy many hold when thinking of AI, AI companies like CrowdStrike are leading the industry in providing risk reduction solutions like intelligent EDR that leverages user and entity behavior analytics (UEBA), an AI-based detection system.
It is no secret that having the right data will transform a business, but when more than 40% of companies do not know where their data is stored, security becomes a secondary issue. Companies like HyperAnna make this a case for automated analytics; AI-powered analytics address security and insight in a single packaged service.
Robotic Process Automation
​
Robotic Process Automation (RPA) might sound complicated, but it is one of the most accessible forms of AI for businesses today, with 53% of enterprises already starting their RPA journey. RPA is simple enough to break into because it starts as small as automating invoicing or backing up files. At its’ current rate, RPA adoption will reach near-universal adoption within five years.
​
Manufacturing
​
The manufacturing world was deeply affected by the global shutdown in 2020. Still, manufacturers battle global supply chain issues, the shortage of raw materials and factory workers, and high turnover resulting in poor product quality and inconsistency issues.
New ways to implement AI that decrease production costs and improve operator efficiency and overall product consistency are emerging quickly. Examples of this are defect detection and predictive maintenance.
​
FinTech
​
Regulations and compliance have always congested the financial world — some so sensitive that the sector is historically slow to adopt new technologies. Even so, 70% of FinTechs already use AI, while 90% use APIs.
Fraud detection technologies are becoming much more sophisticated and with the addition of AI-powered structured data processing and automated audit trails, the financial world’s digital acceleration is increasing at an unforeseen pace.
Sales & Customer Experience
​
AI enables salespeople to view their sales cycles at a 360-degree view; going beyond quantitative analytics and providing qualitative metrics to understand what is happening in a sales pipeline. AI accelerates growth in the top line from salespeople and the scale of understanding and helpfulness to customers.
​“Consumers aren't dumb — they're very good at detecting automated messaging. But AI is not creative, and it's not going to take over the world anytime soon. What AI is great at is being put into a familiar environment and then making decisions — based on patterns and training — at a rate no human can.”
— Don Bosco, Founder @BHuman.ai
What To Prepare Before Adopting AI
Currently, only 37% of companies employ AI. Adopting AI is not much different from onboarding any other software into a business; instead, what defines a company’s ability to adopt AI is whether a strategy is in place or not. AI companies solve problems, but if the business is not ready to solve those problems, AI will not be of any assistance.
Here is what AIP’s Affiliate companies suggest to get started adopting AI today:
​
1. Data Infrastructure
​
The ability to scale data storage is critical to the success of onboarding a new AI service. The significant speed and power needed alone are enough to consider. Without having the proper data infrastructure, such as a data warehouse or lake, set up in advance, the discovery phase will fail before it even begins.
“Anything that can be done with AI to optimize storage usage, extract or summarize data so that the source can be discarded, compress data, or even reduce the need for data collection will be beneficial to current storage patterns and solutions.”
— Chris Andrews, COO & Head of Product @Rendered.ai
2. Strategic Goals
​
Business goals should pilot technology decisions. AI is not a silver bullet solution; it requires significant human interaction upfront to see the long-term ROI. This kind of unpredictable ROI is hard for many businesses to swallow. Nevertheless, that is exactly why strategic KPIs are vital to the success of the adoption.
3. Investable Resources
The discovery phase of adopting AI is not straightforward. With most software these days, downloading an onboarding Wizard and clicking through the steps is the norm. For AI, the typical discovery phase is 4-6 steps, with each requiring investment along the way.
“Most people learn about AI from media and movies, so misinformation spreads fast. The time it takes to cultivate the right dataset, skillset, and mindset for AI leads to a longer discovery phase.”
— Monish Gandhi, CEO & Founder @Gradient Ascent
The Future is AI
It’s clear to see that the future of AI is still unknown; it's a fuzzy crystal ball at best. However, what we do know is that we can all begin learning about the emerging technologies and how they interact with us all on a day-to-day basis.
There's no denying AI’s inevitable integration. What we do with that integration is still up to us.
“Companies are looking for new ways to implement cutting-edge technologies such as AI that decrease production costs and improve operator efficiency and overall product consistency.”
— Sheldon Fernandez, CEO @Darwin AI
For Business Owners
The time to start the conversation around AI adoption is today. Most leaders will likely face pushback within their organizations, but the conversation must be had one way or another.
By starting now, businesses can ensure their future success. Lack of knowledge of future technology does not slow adoption; it leaves stubborn businesses behind.
For Investors
​
After C3.ai went IPO, AI stocks have been on many investors' minds. AI could lead to an economic boost of a whopping $14 trillion in GVA by 2035, according to Accenture. However, as previously noted, some investors feel AI is still too risky at this stage.
​
For AI Companies
​
The competition will continue to rise as VC funding grows and technology advances. Much like the tech startup boom in the early 2000s, AI startups face a lot of unknowns — which allows anyone to pave the way. Pay attention to companies connecting AI startups, and end-users like we are.
​
For Everybody Else
​
AI is the future, and staying cognizant is an active part of staying nimble to that future. By paying attention to trends early on, recognizing opportunities earlier, and watching how the industry develops alongside emerging technologies, we can all be better prepared to adopt the kind of AI we want in our lives — while leaving the rest to develop further.
​
Written by Atlas Katari + AIP Team
Sources:
​
Youn, S. (2018, October 26). Christie's auctions the first AI-produced art, and the gavel price was staggering. ABC News. Retrieved January 21, 2022, from https://abcnews.go.com/Business/christies-sells-ai-produced-art-half-million-dollars/story?id=58749667
​
Blender Bot 2.0: An open source chatbot that builds long-term memory and searches the internet. Meta AI. (2021, July 16). Retrieved January 21, 2022, from https://ai.facebook.com/blog/blender-bot-2-an-open-source-chatbot-that-builds-long-term-memory-and-searches-the-internet/
​
Jia, Y., & Weiss, R. (2019, May 15). Introducing Translatotron: An end-to-end speech-to-speech translation model. Google AI Blog. Retrieved January 21, 2022, from https://ai.googleblog.com/2019/05/introducing-translatotron-end-to-end.html
​
Somers, M. (2020, July 21). Deepfakes, explained. MIT Sloan. Retrieved January 21, 2022, from https://mitsloan.mit.edu/ideas-made-to-matter/deepfakes-explained
​
Wolf, P. G. M. (2022, January 5). Legislation Related to Artificial Intelligence. Legislation related to Artificial Intelligence. Retrieved January 21, 2022, from https://www.ncsl.org/research/telecommunications-and-information-technology/2020-legislation-related-to-artificial-intelligence.aspx
​
Jobs of the future index-future of work: Cognizant. Cognizant Technology Solutions. (n.d.). Retrieved January 21, 2022, from https://www.cognizant.com/jobs-of-the-future-index
​
Rimol, M. (2021, September 7). Gartner identifies four trends driving near-term Artificial Intelligence Innovation. Gartner. Retrieved January 21, 2022, from https://www.gartner.com/en/newsroom/press-releases/2021-09-07-gartner-identifies-four-trends-driving-near-term-artificial-intelligence-innovation
​
Astrid Schepman, Paul Rodway, Initial validation of the general attitudes towards Artificial Intelligence Scale, Computers in Human Behavior Reports, Volume 1, 2020, 100014, ISSN 2451-9588, https://doi.org/10.1016/j.chbr.2020.100014. (https://www.sciencedirect.com/science/article/pii/S2451958820300142)
​
Futurism. (2017, March 20). Google created an AI that can learn almost as fast as a human. Futurism. Retrieved January 21, 2022, from https://futurism.com/google-created-an-ai-that-can-learn-almost-as-fast-as-a-human
​
Moore, S. (2018, January 18). 4 steps to overcome data quality challenges. Gartner. Retrieved January 21, 2022, from https://www.gartner.com/smarterwithgartner/how-to-stop-data-quality-undermining-your-business
​
The state of AI in 2020. McKinsey & Company. (2020, November 27). Retrieved January 21, 2022, from https://www.mckinsey.com/business-functions/mckinsey-analytics/our-insights/global-survey-the-state-of-ai-in-2020#
​
Hazell, W. (2021, February 3). Number of students studying AI at university jumps 400% - but languages slump. inews.co.uk. Retrieved January 21, 2022, from https://inews.co.uk/news/education/students-studying-ai-university-jumps-languages-slump-856740
​
Wiggers, K. (2021, September 30). VCs invested over $75B in AI startups in 2020. VentureBeat. Retrieved January 21, 2022, from https://venturebeat.com/2021/09/30/vcs-invested-over-75b-in-ai-startups-in-2020/
​
Peng T. Ong, M. H. V. (2018, July 21). Why I don't invest in Ai. VentureBeat. Retrieved January 21, 2022, from https://venturebeat.com/2018/07/21/why-i-dont-invest-in-ai/
​
Tripathi, P. (2020, September 6). What vcs look for before investing in an AI startup . VCBay News. Retrieved January 21, 2022, from https://www.vcbay.news/2020/08/23/what-vcs-look-for-before-investing-in-an-ai-startup1/
​
White, P. (n.d.). Artificial Intelligence: Localization Winners, Losers, Heroes, Spectators, and You. Pactera and Nimdzi Insights.
​
Minetti, S. (2020, December 18). What are the main obstacles that are preventing AI startups from scaling up? – thought leaders. Unite.AI. Retrieved January 21, 2022, from https://www.unite.ai/what-are-the-main-obstacles-that-are-preventing-ai-startups-from-scaling-up-thought-leaders/
​
Beauhurst. (2020, September 17). Analysing the impact of covid-19 on UK Ai Startups. Beauhurst. Retrieved January 21, 2022, from https://www.beauhurst.com/blog/analysing-the-impact-of-covid-19-on-uk-ai-startups/
​
Hyper Anna. (n.d.). The Anatomy of Automated Analytics.
​
Maguire, J., & Ehrlich, C. (2021, December 22). Top performing artificial intelligence (AI) companies of 2022. Datamation. Retrieved January 21, 2022, from https://www.datamation.com/artificial-
intelligence/ai-companies/
​
https://www.gocoder.one/blog/active-machine-learning-communities
https://anatomyof.ai/index.html
https://www.nature.com/articles/s42256-020-0219-9
https://arxiv.org/abs/1906.02243v1
https://news.mit.edu/2020/artificial-intelligence-ai-carbon-footprint-0423
https://towardsdatascience.com/most-businesses-dont-use-data-the-right-way-research-e1dcc81a5627
https://hbr.org/2021/09/ai-adoption-skyrocketed-over-the-last-18-months
https://www.beauhurst.com/blog/analysing-the-impact-of-covid-19-on-uk-ai-startups/
https://www.lepide.com/blog/more-than-forty-of-companies-dont-know-where-their-data-is-stored/
https://www.worldfinance.com/markets/the-future-role-of-ai-in-finance
https://link.springer.com/article/10.1007/s11747-019-00696-0
https://www.esa.int/Applications/Observing_the_Earth/Working_towards_AI_and_Earth_observation