From online dating to cybersecurity, AI is routinely working behind the scenes in various aspects of our day-to-day lives.
From smart infrastructure grids to bot-authored news reports, algorithms and artificial intelligence capabilities are routinely working behind the scenes in various aspects of our day-to-day lives. COVID-19 only accelerated the adoption of automation across industries and Gartner pegged “smarter, responsible [and] scalable AI” as one of its top 2021 data and analytics tech trends. In this roundup, we’ve highlighted some of the ways AI is transforming everything from animal conversation efforts to matchmaking in the digital age.
The agtech company AppHarvest is using a number of transformative practices to reimagine farming in the 21st century, including AI. The company is tapping computer vision and AI to help its robo-harvester, Virgo, pick ripe produce right from the vine.
The robotic harvester uses a suite of cameras and infrared laser to map its work environment and uses this information to assess a tomato’s orientation and gauge whether it is “ripe enough to pick,” a company press release said. These scans allow Virgo to determine the “least obstructive and fastest route” to pluck produce using its onboard gripper and arm.
In August, the produce-picking harvesting robot flexed new dexterity skills as it picked strawberries and cucumbers. (Previously, Virgo was shown picking tomatoes off the vine in other videos.)
“With robots roving through the facility, interacting with and caring for the crops, we will be continuously collecting data on plant production to feed into AI and then using software to align facility operations with sales and logistics, making farming as reliable and predictable as a factory,” said Webb at a recent AppHarvest earnings meeting.
SEE: IT expense reimbursement policy (TechRepublic Premium)
Automation and computer algorithms could also transform the roles of humans in the financial services sector. As we reported earlier this year, “robots” were trusted more than people when it comes to money management, according to an Oracle study published in February. Overall, the vast majority of respondent business leaders (85%) wanted “help from robots for finance tasks” and about half (56%) believed robots would “replace corporate finance professionals” in the next half-decade, according to the study.
“AI and machine learning are becoming more prolific in nearly every area of the banking sector, from back office applications and customer engagement through to compliance,” said Jason Somrak, chief of product and strategy at Oracle Financial Services.
Specifically noting financial crime and anti-money laundering, Somrak said these are areas where these applications “are having a tremendous impact.”
“While traditional rules-based AML scenarios may keep financial institutions technically compliant, they are unable to adapt to the constantly changing patterns of today’s criminals,” Somrak said. “However, more are starting to leverage the technology to identify a criminal’s digital ‘fingerprint.'”
Using historical and current data, Somrak said machine learning algorithms are constantly learning and this helps “identify recurring or shifting criminal behavior patterns” to “connect suspicious money movements between criminal organizations.” On the “emerging AI” front, Somrak discussed “deploying smart artificial agents” to help identify gaps in an organization’s compliance controls.
SEE: Juggling remote work with kids’ education is a mammoth task. Here’s how employers can help (free PDF) (TechRepublic)
Researchers around the globe are tapping a vast suite of technologies to help with wildlife and land conservation efforts. This includes using facial recognition to track bears in Canada to deep-learning enabled wildfire prediction. As we reported in December, a team of researchers in Australia are using AI-equipped drones to protect the iconic koala from habitat destruction and more.
To increase the efficiency and accuracy of koala counts, Grant Hamilton, associate professor of ecology at QUT, and his team developed a methodology that uses drones, thermal cameras and AI. But why is it important to count the koalas?
“How do we know that our management actions are having any effect at all? Well, we have to be able to count [the koalas]. So, counting these threatened species is fundamental to making sure that we preserve them. Unless we can do it accurately though and efficiently, it’s not going to get done, and that’s the problem at the moment,” Hamilton said.
When we spoke with Hamilton about the program in December, he estimated that a four-person research team could cover approximately 10 hectares per diem, and the drone-enabled AI method allows the team to cover 50 hectares in two hours.
In recent weeks, a number of high-profile cyberattacks have reverberated across critical U.S. infrastructure ranging from petroleum production and meat manufacturing to local water supplies. As we reported earlier this year, a number of teams overseeing network security at water treatment facilities are using AI-enabled systems to provide round-the-clock monitoring and response for short-handed IT teams.
But the increasingly common use of “artificial intelligence” at times necessitates a more specified semantic conversation. After all, are these applications truly AI or are these solutions more akin to pattern matching?
“From my perspective, artificial Intelligence is a general term that refers to a software with a specific set of goals. In general, most of the current techniques used by security teams are better defined as machine learning algorithms,” said Peleus Uhley, lead security strategist and principal scientist at Adobe. “Our team frequently uses machine learning algorithms to solve various problems in computer security, one example is anomaly detection.”
Using machine learning to detect anomalies is “distinctly different from pattern matching,” Uhley said, adding that “you know in advance what you would consider to be an anomaly” when using pattern matching; meaning teams “need to have a pre-defined, fixed set of patterns that you are matching against for a given environment.”
However, machine learning algorithms allow teams to “take a more generalized approach” and apply the same algorithm across a number of environments, he added.
“A machine learning algorithm is “taught” what is “normal” for each given environment and can then identify anomalies from that baseline. This can often produce better results than pattern matching because the ML algorithm is not limited to a finite set of pre-defined rules,” Uhley said.
“It may be able to detect things that are outside the scope of pattern matching,” he continued.
Earlier this month, Kaspersky published a report about the use of computer algorithms in dating apps and sentiments regarding the role these algorithms play in modern matchmaking. Overall, 44% of respondents “would trust AI or an algorithm to find them a compatible match” and a similar number (43%) prefer “to only see people who have been determined to be a good match by an algorithm,” according to Kaspersky.
Conversely, more than one-third of respondents (39%) said they “find it dehumanizing to be sorted by an algorithm,” 58% would prefer to “have equal access to everyone on an app” rather than having an “algorithm sort people for them,” and more than half (56%) do not believe algorithms “can truly capture the complexity needed to understand attraction,” according to the report.
Algorithms are also being implemented to add a layer of safety to dating apps. Kaspersky security expert Vladislav Tushkanov and security researcher David Jacoby said that machine learning algorithms can help identify bots, potentially identify instances of grooming as well as catfishing and use natural language processing to detect “abusive language or inappropriate messages, such as spam or promotional texts.”
“Computer vision, on the other hand, can automatically filter out unwanted sexual imagery (unless the user actually wants to engage in sexting). Finally, algorithms can be applied to analyze user behavior to block fraudulent accounts,” the Kaspersky representatives explained.
While many apps will tout AI-enabled capabilities with their latest products and services, questions remain about the accuracy of some claims; namely, is this actually artificial intelligence, pattern matching or clever marketing?
“I imagine that many of the uses of technology to make matches would fit definitions of AI that we use. I am sure some of the matches are working off simple heuristics – you’re a match with someone in your ZIP code if there are no other people to be linked to, for example,” said Whit Andrews, distinguished analyst at Gartner.
“I’m sure others are more sophisticated, employing much richer analyses that establish n-dimensional polygons that define a given person, or behavioral matches that align to variables even as far as whether you’re online at the same time,” he said.
To summarize these points, Andrews said that he is “sure” the companies “use AI, but many people would say pattern matching is AI. I am not sure that they always use probabilistic analysis, but I am sure that they do sometimes.”