BY MAX UFBERG FastCompany
8 MINUTE READ
Explore the full 2023 list of Fast Company’s Most Innovative Companies, 540 organizations that are reshaping industries and culture. We’ve selected the firms making the biggest impact across 54 categories, including augmented and virtual reality, consumer electronics, gaming, and more.
We’re witnessing the dawn of the golden age of artificial intelligence—and all the complex questions it brings with it. Over the last few months, tools like OpenAI chatbot ChatGPT and Nvidia’s 2D-to-3D converter NeRF have captured the public’s imagination, prompting a seemingly endless stream of commentary about the industrial, artistic, and ethical implications of generative AI.
Of course, AI is much more than chatbots. And oftentimes, the technology isn’t nearly as eyebrow-raising—and, arguably, scary—as, say, Microsoft Bing’s alter ego Sydney. Take, for example, Voxel, whose monitoring system analyzes video feeds to assess potential workplace threats on factory floors. Or Pano, which uses a sophisticated cloud software and camera network to tip off fire departments on developing disasters. Or Signifyd, which uses machine learning to shield merchants from payment fraud. Or Unlearn’s deployment of digital twins to recreate patient variance in clinical trials. All are examples of AI, and none give off Skynet vibes.
So yes, let’s think long and hard about how this tech will shape our lives, but let’s also acknowledge the wide scope of product offerings, and recalibrate our alarm meters accordingly.
For giving us a glimpse of the future
Founded in 2015, OpenAI calls itself a research and deployment company. On both fronts, it’s been wowing people with its recent releases. The AI image generator Dall-E 2 improves upon its predecessor by offering a higher-resolution text-to-image AI model. In fact, images generated by Dall-E 2 are five times more realistic and accurate than those created with the original model. Even more impressive, Dall-E 2 offers “outpainting,” a new feature that allows users to extend existing images beyond their original borders. For example: Imagine Vermeer’s painting Girl with a Pearl Earring warped to around 20 times its original size. At the same time, the company has continued to enhance its natural-language processor, GPT-3, to process more detailed prompts. (GPT-4, which is reportedly far more powerful, is expected later this year.)
In late November, OpenAI released ChatGPT, an AI-powered program that tries to answer user questions as a human might. ChatGPT made an even bigger splash on the internet than Dall-E 2, with seemingly every Extremely Online person having some fun with the tech, using the bot to produce essay drafts, develop business plans, and even mimic Shakespearean prose. The possibilities with the bot—and Microsoft’s new Bing AI chat, also based on OpenAI technology—may not be limitless: Both have made some notable flubs when it comes to factual accuracy. But whatever’s ahead should be exciting.
Read more about OpenAI, honored as No. 1 on Fast Company’s list of the World’s 50 Most Innovative Companies of 2023.
For putting drug discovery on warp speed
DeepMind‘s AI tool has been tackling one of science’s deepest mysteries: the structure of proteins, which are the building blocks of life. As the Alphabet-owned company explains, “because a protein’s shape is closely linked with its function, knowing a protein’s structure unlocks a greater understanding of what it does and how it works.” A year ago, DeepMind released AlphaFold, an AI system that predicts the 3D structure of a protein from its amino acid sequence, and created the AlphaFold Protein Structure Database to share this knowledge with the scientific community. In 2022, the tool, which has billed itself as the “Google search for protein structures,” added more than 200 million new proteins to its growing database, providing a huge lift to scientists working in drug and vaccine discovery. DeepMind says its findings will contribute to a young field of science called metaproteomics, which deals with experimental study of proteins in microbial contexts. On a less potentially world-changing note, DeepMind also found an algorithm to perform matrix multiplication at record speeds, a testament to its AI’s speed and efficacy.
Read more about DeepMind, honored as No. 27 on Fast Company’s list of the World’s 50 Most Innovative Companies of 2023.
For making mobile app development accessible to everyone
Artificial intelligence powers Builder.ai‘s no-code app development platform, which claims to create software six times faster and 70% cheaper than human teams could. In 2022, the company launched Natasha, a product manager bot, to help guide customers through the app-building process, and cemented a new partnership with JPMorgan Chase to sell Builder products to the financial services giant’s customer bases. Builder.ai raised $100 million in Series C funding in March 2022, bringing its total funding haul to a $195 million.
4. ABNORMAL SECURITY
For protecting companies from targeted email attacks
Using AI to analyze organizational behavior, Abnormal Security is able to detect and prevent email-based cyberattacks like phishing and ransomware. After reaching unicorn status in May 2022 thanks to a $210 million Series C and $4 billion valuation, Abnormal recently launched its new Security Posture Management product, which gives security teams immediate visibility to each of the potential entry and exit points to the cloud email platform. The company already counts Xerox, Groupon, and Royal Caribbean Group among its clients. With the FBI reporting that email attacks cost companies $43 billion in losses between 2016 and 2021, Abnormal’s opportunities look poised to only grow from here.
For creating top-shelf Autonomous Speech Recognition software
B2B startup Speechmatics‘ AI can translate speech to text, regardless of accent or speaking patterns. According to a study published by Stanford University researchers, its AI recorded an overall accuracy of about 83% for Black speakers; Google and Amazon only scored 68%. In December 2022, Speechmatics announced it had begun using Nvidia’s accelerated computing to improve its ML capabilities, a transition that could get the company ahead of the AI pack. That same month, the company announced a partnership with Veritone Voice and Cameo to provide its AI software to the latter’s new “Cameo Kids” feature, registering kids’ dialogue and creating AI voice actors. In June, Speechmatics raised $62 million in a Series B funding round.
For using digital twin technology to create faster clinical trials with smaller control groups
Unlearn uses existing medical data to create digital twins of people—that is, computer-created digital representations of them—that replicate patient variance and characteristics in clinical trials. By reducing control group sizes by up to 35% in phase 2 and 3 trials, the company ‘s tech allows drug companies to conduct smaller (and thus faster) product trials, potentially bringing life-saving medications to market at a much faster pace. Though Unlearn has been around only since 2020, it’s already netted some $70 million in funding (including a $50 million Series B in April) and has been used in trials for neurodegenerative diseases including Alzheimer’s, Parkinson’s, and ALS. In 2022 Unlearn also inked a number of partnerships with pharmaceutical companies, including Merck.
7. PANO AI
For helping fire departments locate blazes before they get out of control
Two-year-old Pano combines AI, sophisticated camera technology, and cloud software to give fire departments immediate insights on burgeoning blazes. Pano’s technology alerts first responders of new fires faster than existing approaches, and its cloud software allows for better coordination between agencies. The company has forged partnerships with six state agencies including ones in California and Montana, and in December 2022 added two states in Australia—one of the most bushfire-prone countries in the world—to its client list. The company, which now monitors a total of 5.6 million acres, raised $20 million in a Series A in September.
For giving e-commerce retailers the tools to detect fraud
Signifyd uses machine learning to protect online merchants from payment fraud, while at the same time shifting liability away from those retailers—a significant offering in an era where more and more shopping is happening online. After considering thousands of variables such as past purchase behavior and location, the company’s software is able to locate elements of the buyer’s data to determine their legitimacy. In addition, vendors have seen a 70% reduction in false declines—that is, legitimate orders that are turned away due to incorrect suspicions—thanks to Signifyd’s software. In 2022, the company teamed with fintech giant FIS to bring its commerce protection tech to the world’s largest payment processor, a significant step in scaling its technology and offering it to the masses. Signifyd’s annual revenue run rate is on pace to increase 40% this year.
For making world building as simple as snapping a few photographs
In 2022, AI computing company Nvidia unveiled Instant NeRF, its AI-powered software that turns 2D photos into a 3D scene in seconds. NeRF has a variety of uses: It can be applied to metaverse tech, can capture video-conference participants in 3D dimensions, and can make lifelike scenes in video games. Users simply have to take a few dozen photos to train the neural radiance fields (or NeRFs); using those pictures, the tech will quickly spit out a finished product. Best of all, it’s free to use for noncommercial purposes; already the tech’s codebase has been downloaded tens of thousands of times. The company’s AI prowess situates it well for benefiting from the current AI boom. As CEO Jensen Huang told investors in February, “We are set to help customers take advantage of breakthroughs in generative AI and large language models.” Nvidia’s 2022 revenue was $27 billion.
For combining AI tools and live video feed to help prevent workplace accidents
Founded in late 2020, Voxel works with industrial and manufacturing businesses to analyze video feeds and determine potential safety threats on the factory floor. Its AI can alert employers to real-time threats (say, a spill), and spot repeated actions that might heighten risk. The technology can even help with things like energy savings by monitoring open doors in refrigerated warehouses. According to Voxel, its tech has led to a 77% reduction of injuries in customers’ workforces and up to a 90% improvement in monitored safety behaviors in the workplace. In 2022, the company raised a $15 Series A and began commercializing its tech.