Heard on the Street – 2/14/2022

Welcome to insideAI News’s “Heard on the Street” round-up column! In this regular feature, we highlight thought-leadership commentaries from members of the big data ecosystem. Each edition covers the trends of the day with compelling perspectives that can provide important insights to give you a competitive advantage in the marketplace. We invite submissions with a focus on our favored technology topics areas: big data, data science, machine learning, AI and deep learning. Enjoy!

Deep Learning is The Way Forward for Understanding Unstructured Data. Commentary by Instabase CEO Anant Bhardwaj

Highly variable, complex documents that require time-intensive and error-prone manual processes make up 80% of an organization’s data. While new technologies over the past few decades such as machine learning have enabled business process automation, the emergence of deep learning techniques is now fundamentally changing how companies across industries approach their document-based processes, drive bottom line efficiencies and transform the customer experience. Unlike the first document understanding solutions, which were template-based or rules-based, today’s deep learning models understand a document’s context and content in its raw form and can process highly variable and complex documents without human intervention. They also become smarter over time, generalizing their learnings across diverse document types and evolving using human-in-the-loop processes. The state-of-the-art deep learning models built based on the transformer architecture deliver more accurate results faster by requiring fewer samples to fine-tune for a specific document use case. Fewer samples, combined with annotation-based model training, enable even non-technical users to utilize the latest deep learning techniques to automate their business processes. Thanks to deep learning, we can now unlock more complex document use cases such as claims, financial documents, and legal contracts, speed up time to delivery, and automate critical processes by bringing diverse talents to the table.

Google Analytics’ resounding privacy implications for US companies. Commentary by Jean-Claude Kuo, Principal Product Manager, Cloud Security, Talend

US technology providers are facing a massive roadblock after a European regulator found Google Analytics services to be illegal. From what I understand, this is not a new ruling, but rather GDPR clauses being applied to a variety of services that recently came under scrutiny from months of Google Analytics examination. Regardless of this battle’s outcome, organizations should continue to adapt and anticipate stricter restrictions on EU Citizen’s PII processing. This development should be seen as a precursor of things to come, as these decisions often trigger ripple effects beyond just the EU and US. Following the CJUE Schrems II ruling that invalidated the EU-US Privacy Shield Framework in 2020, the European Data Protection Board (EDPB) formulated data encryption as an effective technical safeguard to be implemented by organizations in order to protect EU Citizen Privacy. While data encryption is a valid technical approach in theory, organizations will continue to rely heavily on US-based services that require unencrypted data to deliver desired outcomes, such as operational analytics and customer experience. This is a difficult exercise for business leaders, as they are tasked with balancing data privacy (with steep fines and reputational damage if violated) and harnessing the full power of data to remain agile and relevant in the market. Innovation often thrives during challenging times. Data management and governance, in addition to privacy-enhancing technologies, create a path for businesses to strive for success in line with regulatory compliance.

ESG goals/net-zero emissions and federal compliance will require advanced data monitoring deployments. Commentary by Todd Kiehn, SVP Global Product Manager at GTT 

The recent COP 21 led to a new international climate agreement, applicable to all countries, aiming to keep global warming at 1.5°C – 2°C. One result of these initiatives is that governments will start to require businesses to reduce carbon emissions throughout their improving supply chains, cutting emissions, and increasing efficiency. To comply with new environmental, social and governance (ESG) objectives, businesses will need to track, monitor, and measure their data like never before. Today business data is available from every device and at every stage of the journey, from physical to digital and the insights generated by this data will help businesses comply with emerging industry and national ESG mandates. New compliance demands will bring additional network security challenges to an ever-expanding digital perimeter and businesses will need to protect valuable data, employees, and the company’s network from bad actors. Corporations will need to lean on the expertise of managed service providers to design solutions for an increasingly complex hybrid work environment as workers and data move further from the core and closer to edge. 

What the White House Zero Trust Data Protection Strategy Gets Right, and What It Still Needs. Commentary by Chris McLellan, Director of Operations, Data Collaboration Alliance

At its most fundamental level, data privacy is about giving citizens and organizations the ability to control who can access their information, and how. The Biden administration’s recommendation that Chief Data Officers at federal agencies get their houses in order with regard to data categorization with a view to automating that process in the medium-term makes a lot of sense. At the same time, it’s important to recognize that current approaches to data sharing can be best described as “data copying” with the vast majority of the replications happening routinely and invisibly between systems, not people. The inevitable result of so many copies is the steady erosion of the control that data owners have over data quality (including categorization) and access. As a society, we already have strict laws in place to prevent the duplication of our currency, intellectual property, and identities in order to protect their value – the same principle works for data, too. The industry needs new technologies, such as Dataware and standards, including Zero-Copy Integration in Canada that help eliminate copies and give control back to data owners, stewards, and custodians. And part of our core mission at the Data Collaboration Alliance is to accelerate this adoption. The administration should consider meaningful data ownership and control as necessary elements of the national cybersecurity strategy.

Data Mesh and Data Fabric, friends not enemies. Commentary by John Wills, Field CTO at Alation

Amid the ongoing debate between data mesh and data fabric, many companies are asking themselves which is the best choice. The answer is both. Data mesh and data fabric are independent, yet complementary design concepts. According to Gartner, data fabric serves as an integrated layer (fabric) of data and connecting processes. By using technologies to automate continuous discovery, analysis, and reuse of metadata, organizations can overcome different challenges and reduce the error-prone manual efforts that often coincide with data work. Data mesh is about the people, calling for a shift to ensure high quality data in the hands of consumers quickly and efficiently. With data mesh, the people closest to the data are responsible for the health of the data, delivering data as a product, and objective measures such as data quality. Data mesh and data fabric are complementary, not competitive. Together, both a data mesh and data fabric approach to business can help organizations keep data at the forefront. Companies should look to adopt a “meshy fabric” approach, taking the best pieces of each to overcome bottlenecks in data work.

Under pressure from activists & lawmakers, IRS drops plan to use facial recognition. Commentary by Caitlin Seeley George, Campaign Director, Fight for the Future

After activists sounded the alarm and lawmakers on both sides of the aisle expressed outrage and concern, the Treasury department has confirmed that the IRS is dropping its plan to use ID.me’s controversial and invasive facial recognition software. Digital rights group Fight for the Future led the charge against the agency’s use of facial recognition, and along with the Algorithmic Justice League, the Electronic Privacy Information Center, and other groups launched DumpID.me. The IRS’ plan to use facial recognition on people who are just trying to access their tax information online was a profound threat to everyone’s security and civil liberties. We’re glad to see that grassroots activism and backlash from lawmakers and experts has forced the agency to back down,” said Caitlin Seeley George (she/her), campaign director at Fight for the Future. “But several other Federal agencies are still using ID.me’s discriminatory and insecure software, including the Veterans Affairs Administration and Social Security Administration, as well as 30 states that use it on people trying to access unemployment benefits. No one should be coerced into handing over their sensitive biometric information to the government in order to access essential services. The lawmakers who led the charge against the IRS use of this technology should immediately call for an end to other agencies’ contracts, and there should be a full investigation into the Federal government’s use of facial recognition and how it came to spend taxpayer dollars contracting with a company as shady as ID.me.

Big Data and Audience Churn: The Case For Hyper-Personalization. Commentary by Rob Delf, CEO, Fabric

After experiencing a meteoric rise, Netflix’s subscriber growth is now slowing down. The streamer reported last month that it fell short of its initial forecast of 8.5 million new subscribers. The answer to how Netflix, or any of its competitors, can maintain market share is to stop subscriber churn through hyper-personalization. Recent advances in A.I. and machine learning tech can be leveraged to create a profile of each subscriber’s data to deliver content specific to them. The days of relying on segment demographics or ‘types’ to identify the content viewers want to see are over. Rather, streamers should be focused on creating a unified data architecture to build true customer-centricity. It’s not a lack of data. The amount of data collected by streamers and broadcasters on a daily basis is overwhelming. In fact, that’s the problem. Too much data is going to waste because it’s not being analyzed properly to provide useful, actionable information about customer behavior.

How AI is entering the corporate training space. Commentary by Geoff Curless elearning Brothers Senior VP of Sales, Learning Solutions 

In this new virtual world we’re all working in, a wide range of effective communication skills has become more crucial than ever. Training workers in soft skills has become both easier and less time consuming as AI finds its way into the learning and development space. Artificial intelligence can be used to run sentiment analysis on videos to give individuals feedback on how they present themselves using indicators such as eye contact, tone, pace, keywords, logic, articulation, facial expressions and filler words. This automated validation is a key role that AI can take over from managers, freeing them up for other mentorship opportunities. Another way that AI is being utilized in training is through individualized learning experiences. Gone are the days when all workers consume the same content and take the same courses, despite being in different roles, having diverse skill sets and goals. With smart AI in place, workers are given a guided learning path that targets specific gaps in each individual’s knowledge. Not only is AI-guided training more effective for the organization, it better addresses each person’s career goals and aspirations. 

Healthcare organizations find actionable Insights with AI and data analytics. Commentary by Amy Brown, Founder and CEO of Authenticx

Within the last two years of navigating the pandemic, healthcare has accelerated or transformed their digital strategy.  Leveraging tools such as artificial intelligence (AI) and data analytics has moved from the “aspirational roadmap” to business-critical.  Healthcare leaders are focused on technology that drives a patient-centric experience.  Why? Because positive experiences lead to customer retention, conversion and positive brand reputation.  Artificial intelligence and data analytics can help surface customer voices, gain real-time understanding from unstructured conversational data, and reveal actionable insights—all in pursuit of quickly and accurately identifying and meeting patient needs. 

AI in cybersecurity: the good, the bad, the evil. Commentary by Ric Longenecker, CISO at Open Systems

AI and ML have taken the security market by storm over the past 5 years, and now cyber attackers have also begun embracing AI to evade detection and create their own storm. This year, we may see AI increasingly used to attack the models within security software and those outputs used to enable malware to evade detection. Although cybersecurity vendors already integrate AI within their platforms, it is important to recognize that this alone is not enough. To truly combat the threat of bad actors exploiting AI, enterprises must ensure their security providers use AI in combination with the human know-how of security experts to implement clear and repeatable processes.

2022 Safer Internet Day Tips & Tricks. Commentary by Peter Tsai, Head of Technology Insights at Spiceworks

A safe internet doesn’t just require protecting data and users by implementing technological solutions or following security best practices. In an online environment plagued by spam, misinformation, trolls, and extremism, online safety requires a human touch. Digital platforms must invest in robust content moderation teams that actively enforce guidelines to keep online discourse focused and civil. Only when internet safe spaces exist can users feel comfortable forming genuine connections with others by taking part the informational and cultural exchange that is so vital to the flourishing of communities that are truly welcoming to all.

2022 Safer Internet Day Tips & Tricks. Commentary by Brian Spanswick, Chief Information Security Officer and Head of IT at Cohesity

Safer Internet Day is a reminder that while the internet is an empowering resource, it’s also a hotbed of increasingly sophisticated criminal activity where every entry point should be considered a vulnerability. To minimize the threat of ransomware and data exfiltration, organizations must adopt a security-first, next-gen data management solution designed to safeguard their backup infrastructure and better protect internal and consumer data. Employees must also do their part by adhering to password protocols, stay mindful of email safety, and keep their devices updated with antivirus software.

2022 Safer Internet Day Tips & Tricks. Commentary by Daniel Markuson, Digital Privacy Expert at NordVPN

As with every new technology, manufacturers want to jump on the Metaverse train as soon as possible and the rush to compete may result in a lack of oversight, allowing hackers to easily take advantage of consumer vulnerabilities using VR technology. Our new study confirms consumer concerns for their privacy and security in the metaverse. Since children tend to not understand the possible risks, they are the perfect targets for cybercriminals. We strongly suggest that parents read the privacy policies at length and instruct children to create anonymous nicknames on VR multiplayer games and to keep their personal information private.

2022 Safer Internet Day Tips & Tricks. Commentary by Gabriel Engel, CEO of Rocket.Chat

When it comes to protecting collaborative data between different universities or universities and businesses, best practices call for intense regulation of data sharing via contracts that specify security policies, data protection responsibilities, technology to be used, and more. Universities are often underfunded and lack resources to acquire and maintain new digital tools. Luckily, more and more governing bodies understand the need for digital transformation in the education sector.

2022 Safer Internet Day Tips & Tricks. Commentary by David McKenna, Senior VP Engineering at Axway

Keep systems up to date. The Internet never forgets – be careful with what you post and the content of what you post. You could leak personal information which can be used to exploit you. Or you could post damaging comments that will haunt you in later life. Strong & unique passwords – machines can brute/guess ‘human’ friendly passwords. Never trust a provider to store your credentials safely, so assume they will be hacked and your credentials (user name and password) will be public someday.

Sign up for the free insideAI News newsletter.

Join us on Twitter: @InsideBigData1 – https://twitter.com/InsideBigData1