Connect with us

Tech News

The paradox of CAPTCHAs: Too smart for humans vs AI?

(TECH NEWS) AI is catching up to our cybersecurity technology and often tricking humans too — so what’s next for CAPTCHAs and the internet?

Published

on

Person using phone with laptop to verify CAPTCHAs and code.

We’ve all encountered it before: The occasional robot test that feels impossible to beat. If you’ve felt like these tests, also known as CAPTCHAs, have gotten harder in the last couple of years, you aren’t wrong—and the reason is as ironic as it is baffling.

Simply put, AI are just as good as—and often better than—humans at completing CAPTCHAs in their classic format. As machine learning and AI become more advanced, the fundamental human attributes that make consistent CAPTCHA formats possible become less impactful, raising the question of how to determine the difference between AI and humans in the future.

The biggest barrier to universal CAPTCHA doctrine is purely cultural. Humans may share experiences across the board, but such experiences are typically basic enough to fall victim to the same machine learning which has rendered lower-level CAPTCHAs moot. Adding a cultural component to CAPTCHAs could prevent AI from bypassing them, but it also might prevent some humans from understanding the objective.

Therein lies the root of the CAPTCHA paradox. Humans are far more diverse than any one test can possibly account for, and what they do have in common is also shared by—you guessed it—AI. To create a truly AI-proof test would be to alienate a notable portion of human users by virtue of lived experience. The irony is palpable, but one can only imagine the sheer frustration developers are going through in attempting to address this problem.

But all isn’t lost. While litmus tests such as determining the number of traffic cones in a plaza or checking off squares with bicycles (but not unicycles, you fool) may be beatable by machines, some experts posit that “human entropy” is almost impossible to mimic—and, thus, a viable solution to the CAPTCHA paradox.

“A real human being doesn’t have very good control over their own motor functions, and so they can’t move the mouse the same way more than once over multiple interactions,” says Shuman Ghosemajumder, a former click fraud expert from Google. While AI could attempt to feign this same level of “entropy”, the odds of a successful attempt appear low.

Jack Lloyd has a BA in Creative Writing from Forest Grove's Pacific University; he spends his writing days using his degree to pursue semicolons, freelance writing and editing, oxford commas, and enough coffee to kill a bear. His infatuation with rain is matched only by his dry sense of humor.

Tech News

4 ways startups prove their investment in upcoming technology trends

(TECH NEWS) Want to see into the future? Just take a look at what technology the tech field is exploring and investing in today — that’s the stuff that will make up the world of tomorrow.

Published

on

Woman testing VR technology

Big companies scout like for small ones that have proven ideas and prototypes, rather than take the initial risk on themselves. So startups have to stay ahead of technology by their very nature, in order to be stand-out candidates when selling their ideas to investors.

Innovation Leader, in partnership with KPMG LLP, recently conducted a study that sheds light onto the bleeding edge of tech: The technologies that the biggest companies are most interested in building right now.

The study asked its respondents to group 16 technologies into four categorical buckets, which Innovation Leader CEO Scott Kirsner refers to as “commitment level.”

The highest commitment level, “in-market or accelerating investment,” basically means that technology is already mainstream. For optimum tech-clairvoyance, keep your eyes on the technologies which land in the middle of the ranking.

“Investing or piloting” represents the second-highest commitment level – that means they have offerings that are approaching market-readiness.

The standout in this category is Advanced Analytics. That’s a pretty vague title, but it generally refers to the automated interpretation and prediction on data sets, and has overlap with Machine learning.

Wearables, on the other hand, are self explanatory. From smart watches to location trackers for children, these devices often pick up on input from the body, such heart rate.

The “Internet of Things” is finding new and improved ways to embed sensor and network capabilities into objects within the home, the workplace, and the world at large. (Hopefully that doesn’t mean anyone’s out there trying to reinvent Juicero, though.)

Collaboration tools and cloud computing also land on this list. That’s no shock, given the continuous pandemic.

The next tier is “learning and exploring”— that represents lower commitment, but a high level of curiosity. These technologies will take a longer time to become common, but only because they have an abundance of unexplored potential.

Blockchain was the highest ranked under this category. Not surprising, considering it’s the OG of making people go “wait, what?”

Augmented & virtual reality has been hyped up particularly hard recently and is in high demand (again, due to the pandemic forcing us to seek new ways to interact without human contact.)

And notably, AI & machine learning appears on rankings for both second and third commitment levels, indicating it’s possibly in transition between these categories.

The lowest level is “not exploring or investing,” which represents little to no interest.

Quantum computing is the standout selection for this category of technology. But there’s reason to believe that it, too, is just waiting for the right breakthroughs to happen.

Continue Reading

Tech News

Internet of Things and deep learning: How your devices are getting smarter

(TECH NEWS) The latest neural network from Massachusetts Institute of Technology shows a great bound forward for deep learning and the “Internet of Things.”

Published

on

Woman using smart phone to control other devices in home, connected to deep learning networks

The deep learning that modifies your social media and gives you Google search results is coming to your thermostat.

Researchers at the Massachusetts Institute of Technology (MIT) have developed a deep learning system of neural networks that can be used in the “Internet of Things” (IoT). Named MCUNet, the system designs small neural networks that allow for previously unseen speed and accuracy for deep learning on IoT devices. Benefits of the system include energy savings and improved data security for devices.

Created in the early 1980s, the IoT is essentially a large group of everyday household objects that have become increasingly connected through the internet. They include smart fridges, wearable heart monitors, thermostats, and other “smart” devices. These gadgets run on microcontrollers, or computer chips with no processing system, that have very little processing power and memory. This has traditionally made it hard for deep learning to occur on IoT devices.

“How do we deploy neural nets directly on these tiny devices? It’s a new research area that’s getting very hot,” said Song Han, Assistant Professor of Computer Science at MIT who is a part of the project, “Companies like Google and ARM are all working in this direction.”

In order to achieve deep learning for IoT connected machines, Han’s group designed two specific components. The first is TinyEngine, an inference engine that directs resource management similar to an operating system would. The other is Tiny NAS, a neural architecture search algorithm. For those not well-versed in such technical terms, think of these things like a mini Windows 10 and machine learning for that smart fridge you own.

The results of these new components are promising. According to Han, MCUNet could become the new industry standard, stating that “It has huge potential.” He envisions the system has one that could help smartwatches not just monitor heartbeat and blood pressure but help analyze and explain to users what that means. It could also lead to making IoT devices far more secure than they are currently.

“A key advantage is preserving privacy,” says Han. “You don’t need to transmit the data to the cloud.”

It will still be a while until we see smart devices with deep learning capabilities, but it is all but inevitable at this point—the future we’ve all heard about is definitely on the horizon.

Continue Reading

Tech News

Google is giving back some privacy control? (You read that right)

(TECH NEWS) In a bizarre twist, Google is giving you the option to opt out of data collection – for real this time.

Published

on

Open laptop on desk, open to map privacy options

It’s strange to hear “Google” and “privacy” in the same sentence without “concerns” following along, yet here we are. In a twist that’s definitely not related to various controversies involving the tech company, Google is giving back some control over data sharing—even if it isn’t much.

Starting soon, you will be able to opt out of Google’s data-reliant “smart” features (Smart Compose and Smart Reply) across the G-Suite of pertinent products: Gmail, Chat, and Meet. Opting out would, in this case, prevent Google from using your data to formulate responses based on your previous activity; it would also turn off the “smart” features.

One might observe that users have had the option to turn off “smart” features before, but doing so didn’t disable Google’s data collection—just the features themselves. For Google to include the option to opt out of data collection completely is relatively unprecedented—and perhaps exactly what people have been clamoring for on the heels of recent lawsuits against the tech giant.

In addition to being able to close off “smart” features, Google will also allow you to opt out of data collection for things like the Google Assistant, Google Maps, and other Google-related services that lean into your Gmail Inbox, Meet, and Chat activity. Since Google knowing what your favorite restaurant is or when to recommend tickets to you can be unnerving, this is a welcome change of pace.

Keep in mind that opting out of data collection for “smart” features will automatically disable other “smart” options from Google, including those Assistant reminders and customized Maps. At the time of this writing, Google has made it clear that you can’t opt out of one and keep the other—while you can go back and toggle on data collection again, you won’t be able to use these features without Google analyzing your Meet, Chat, and Gmail contents and behavior.

It will be interesting to see what the short-term ramifications of this decision are. If Google stops collecting data for a small period of time at your request and then you turn back on the “smart” features that use said data, will the predictive text and suggestions suffer? Only time will tell. For now, keep an eye out for this updated privacy option—it should be rolling out in the next few weeks.

Continue Reading

Our Great Partners

The
American Genius
news neatly in your inbox

Subscribe to our mailing list for news sent straight to your email inbox.

Emerging Stories

Get The American Genius
neatly in your inbox

Subscribe to get business and tech updates, breaking stories, and more!