Welcome to Thomas Insights — every day, we publish the latest news and analysis to keep our readers up to date on what’s happening in industry. Sign up here to get the day’s top stories delivered straight to your inbox.

Thoughts After Dark answers the questions you have in the final moments before drifting off to sleep when a simple Google search turns into an hour-long exploration into how things are made and how they work. Your random late-night questions are answered here — even the ones you didn’t know you had.

“Is that a selfie?” I’ve asked myself this question a dozen times over the past few weeks as I scrolled Instagram and Twitter and noticed the people I follow turning themselves into Renaissance-style paintings and pastel fairies. 

Alongside these semi-realistic portraits on my social media feed live beautiful mid-century modern mansions with underground hot tubs surrounded by enticing steam. It took me a while to realize that none of these homes actually existed and that the friends and celebrities I followed turned themselves into a piece of art with the click of a button (and $3.99). 

Plus, a lot of help from artificial intelligence (AI) technology. 

When I first heard of AI, I couldn’t help but envision Ava and Samantha from the films Ex Machina and Her, respectively. Now, it feels like every other headline I read mentions new innovations in AI. But the technology hasn't only infiltrated social media feeds, it’s being used in industries from healthcare to manufacturing. 

As Ryan Broderick put it in his newsletter Garbage Day earlier this month, “The last nine months of A.I. have felt like a speedrun of the last 15.” A 2021 report from KPMG that surveyed 950 business and government decision-makers also found that about half of the people surveyed thought AI adoption was moving too fast, especially in industries like industrial manufacturing, retail, and tech. 

Amid the widespread adoption on social media and robust excitement for what AI can accomplish in industry, how valid are the concerns about the rapid development of AI tech?

Why Is AI Evolving So Fast? 

The pandemic shifted the business strategy of many sectors and industry leaders and inspired companies to amp up their digital transformation by implementing data analytics and AI tools. A survey by PwC, for example, found that over 50% of companies accelerated their AI adoption amid the pandemic, and 86% noted that it became a “mainstream technology” within their organization in 2021. 

But why? The pandemic quickly shifted work arrangements and consumer purchasing habits, in turn increasing the need for access to certain data. AI gives greater insight into who customers are, supply chains, and business models. 

“[CEOs] know they need to build a data foundation, taking all of the data sets, putting them into an insights engine using all the algorithms, and powering insights solutions that can help them optimize their businesses,” Arnab Chakraborty, global managing director at Accenture, told the Harvard Business Review.

The Concerns About the Rapid Adoption of AI 

Currently, the most popular iPhone app in the U.S. is Lensa AI. The app uses Stable Diffusion, a powerful AI-based image generator created by OpenAI, to create striking images based on an inputted image and text. It’s also the culprit for the AI-generated “selfies” that continue to take over my social feed. 

Since the beginnings of Instagram, Twitter, Facebook/Meta, and now TikTok, photos and videos have been taken and reposted, often without crediting the original creator. This has become of increasing concern with the growing popularity of AI systems, especially regarding art. It has raised questions about what is considered “stealing” and if the technology will eliminate jobs. 

As systems created by research laboratories like OpenAI continue to gain popularity, there are a few concerns about the rapid adoption of AI technology. But there are also many benefits. 

Security Concerns 

A report from Deloitte discovered that the number one risk that businesses associated with adopting AI was cybersecurity vulnerability. 

Because AI and machine learning systems require higher volumes of complex data, systems are vulnerable to exploitation. This can happen through manipulation, in which the system is given malicious inputs to make faulty predictions based on unverified data.

Data privacy is also of concern, as AI systems rely on large volumes of datasets, and any malfunction or hack could be detrimental to security and output.

AI Can Improve Cybersecurity  

While security is a concern, AI can actually be used to improve cybersecurity efforts. 

The rise in cyberattacks has fueled AI-based security products. By 2030, the market is anticipated to hit $133.8 billion. 

While AI can be used for data loss prevention, antivirus software, and fraud detection, there aren’t many businesses taking advantage of it. Instead, they are using AI for behavior analysis to uncover patterns by hackers like timing, method, and how they move within a system.

Gartner interviewed 50 security vendors and discovered that a majority were using AI to filter signals versus actual attacks to prioritize responses. AI threat monitoring also eliminates repetitive tasks for employees and can pinpoint potential vulnerabilities in a system.   

Stolen Intellectual Property

In August, author Stephen King shared an image of Pennywise — the clown from his acclaimed novel It — riding a bike. It’s perfectly spooky, but the signature in the bottom right corner tells an interesting story: the AI system used to create the art pulled from millions of pieces of artwork online. 

And in September, Meta announced a new feature that allows users to create a video by simply typing what they want to see. However, the system was reportedly trained unsupervised and uses “unlabeled video content” or nonspecific content online. 

These occurrences led to the creation of the websites Have I Been Trained? and Lexica, which allows artists to search if their work is in the library of images used to train Stable Diffusion and Google Imagen, another text-to-image system.  

Grzegorz Rutkowski, a gaming illustrator that worked on Dungeons Dragons, found out his work has been used over 90,000 times by Stable Diffusion through Lexica. Because AI systems are often trained using copyrighted images without permission, it raises the question of whether the art is borrowed or stolen. 

AI Can Make Technology More Accessible 

On the other hand, Prisma Labs CEO Andrey Usoltsev told NBC News that these AI systems are democratizing access to technology. “As AI technology becomes increasingly more sophisticated and accessible, it is likely that we will see AI-powered tools and features being widely integrated into consumer-facing apps at a rapid scale,” Usoltsev said. 

Now, she notes, technology that wasn’t always available to the untrained is accessible for all to enjoy. 

Additionally, artists may turn to AI for creative inspiration if reference images aren’t already available online or to visualize a scene in a movie or screenplay. 

In November, the San Fransisco Ballet utilized Midjourney, a text-to-image system, to promote its upcoming performance of “The Nutcracker.” While some artist advocacy groups criticized using the AI-powered system, a spokesperson for the ballet said it was a “chance to experiment with today's technological tools.”

Ethical Concerns

Because AI systems learn based on inputs, there is potential for biases. Bias occurs when two data sets in a machine learning system aren’t considered “equal,” in turn skewing the data in favor of one. 

In 2018, for example, Amazon had to scrap an AI-recruiting tool that showed bias against women. Automation was key to the e-commerce giants recruiting efforts, who were searching for quicker ways to find top talent. The experimental tool rated applicants from one to five. But because the tool was trained by observing patterns in resumes submitted over a 10-year period, and most of these applications came from men, the system penalized resumes that included the word “women’s” and favored male applications. 

Paving the Way for Improvement 

AI is helping make businesses smarter and more efficient and has paved the way for a deeper understanding of data. 

Teams of dedicated data scientists have tuned into figuring out how AI bias works and how to prevent it. Dror Zaifman, director of digital marketing for iCASH, for example, said that his team trains AI to be “prescriptive but always transparent.” 

AI technology can also help eliminate human bias, as “a human-built scoring model is subject to the biased beliefs of its developers.”

Because eliminating bias also relies on understanding users, AI-powered systems have prompted more customer conversations. This means walking through the AI experience with customers to understand what’s working and what isn’t. 

The Future of AI

According to a survey by The AI Journal, 72% of business leaders feel positive about the future role of AI. More specifically, nearly 75% believe the technology will power more efficient business processes. It’s boosting productivity, delivering new services, addressing supply chain strains, and making predictions to prevent disruption. It’s also helping address labor shortages.  

Last year, “The Great Resignation” became a part of the zeitgeist. It filled headlines and impacted industries from hospitality to logistics. But AI is helping fill the skills gap through employee support tools, helping predict future talent requirements, and making data-driven business decisions. 

In the future, AI will continue to optimize business processes across industries from healthcare to hospitality. For now, you can learn more about how the digital manufacturing marketplace Xometry uses AI to power its Instant Quoting Engine® and connect buyers.  

It might take some time to get used to seeing AI-generated portraits on Instagram and houses that only exist on Twitter, but AI isn’t going anywhere.  

More Thoughts After Dark: 

• How VR and AR Are Infiltrating Industries from Aviation to Fashion
• 5 Billion Views and Counting: Why Are Factory TikToks Trending?
• How Shapes Help Us Decipher Which Movie Character Is the Villain and What Products to Buy

Image Credit: enzozo / Shutterstock.com

More from Technology