The quarterly publication of the International Legal Technology Association
Issue link: https://epubs.iltanet.org/i/1172342
P E E R T O P E E R : I L T A ' S Q U A R T E R L Y M A G A Z I N E | F A L L 2 0 1 9 27 categorize it as negative because of other data that's factored in. The technolo also takes into account behavior that could be considered anomalous, such as communication happening after normal business hours, or two people communicating frequently during a certain stand-alone period of weeks or months. It might also consider what subjects were discussed during those communications and whether they seemed atypical compared to what those same people discuss with others — or even if their words were chosen to conceal the true nature of their conversation. The technolo also factors in who is talking to whom and what relevant events might be taking place in the time period when the messages are being sent. With the help of this AI-provided context, the data suddenly gains a new dimension, and documents that might not have seemed important before rise to the top of search results. Unlike other forms of AI, emotionally intelligent AI requires no training of a system. It can be incorporated at the very beginning of the eDiscovery process and works nearly instantaneously. It can help law firms gain more accurate insight into a set of data, enable them to eliminate non-responsive documents early on, and allow for human review to be more focused as the data moves through the eDiscovery process. incorporate modifiers like OR, AND and NOT, and proximity searches, which produce results based on the proximity of search terms to each other). While these methods are useful in many situations, their ability to filter down data is limited. These so-called "static" searches examine words as a series of letters without considering their meaning or context. But emotionally intelligent AI adds an additional layer of comprehension in which words are read, not just seen. If you listen to the industry chatter, you may have heard such ludicrous claims as: "Keyword searching is dead!" Well, not quite, but our prediction is that in the not-too-distant future, emotional categorization will be right up there with keyword searching and date range filtering when we're scoping out search parameters. Emotionally intelligent AI searches documents for language that conveys a range of emotional states and situations, such as positivity or negativity, as well as intent, opportunity, pressure and rationalization. Part of this is conveyed through the words themselves — derogatory words, for example — but also in the way sentences are phrased and punctuated (the frequency of exclamation points or capital letters used, for example). The AI puts these indicators and others together to make a determination. It can be surprisingly perceptive — a document might not contain any bad language or other telltale signs of negativity, yet the system might still It can be surprisingly perceptive — a document might not contain any bad language or other telltale signs of negativity, yet the system might still categorize it as negative because of other data that's factored in. WHAT IS A DATA BREACH? The unauthorized access, use or retrieval of data by an individual, group, application or service. 2,216 data breaches reported worldwide in 2018 28% of attacks in 2018 involved insiders 61% of breaches hit smaller businesses in 2018 $3.86 million average cost of a data breach Number of records exposed per year 0 100 200 300 400 500 2014 2015 2016 2017 2018 Ransomware Accidental exposure to the internet Unauthorized access Data theft by an insider Employee negligence or error Phishing Physical theft or exposure of data Data Breach Discovery ™ : How to Identify and Report on Compromised Data A DATA BREACH CAN HAPPEN IN A VARIETY OF WAYS: ERROR Cyber attack or hacking 1 2 3 4 5 6 7 8 ! • Identifying and collecting compromised data sources • Searching the huge amount of emails and files that are often exposed in a breach • Processing and filtering data with sophisticated analytics and leading-edge technology • Employing eDiscovery principles, streamlined workflows, forensic investigation techniques, experienced document reviewers and a detailed understanding of privacy issues and industry regulations • Reviewing data to validate findings and ensure quality control • Capturing affected names and key data • Creating a unique list of affected individuals • Providing the details you need to resolve a data breach and sleep easier Fast, accurate, effective identification and reporting of compromised personally identifiable information and sensitive data 126% increase in the number of exposed records with personally identifiable information (PII) in 2018 $148 average cost per lost or stolen record in 2018 Millions of records exposed (888) 338-4242 | biaprotect.com Sources: • Verizon 2018 Data Breach Investigation Report • 2018 Cost of Data Breach study by Ponemon Institute • Statista • Identity Theft Resource Center • IT Governance Blog • Chief Executive 85.61 million 169.07 million 36.6 million 178.96 million 446.52 million WHAT IS DATA BREACH DISCOVERY™? The exposure of just 25 email inboxes can yield over 1 million documents. RESOLVING A DATA BREACH When a data breach occurs, companies need to resolve the problem quickly by: • Securing any exposed emails and loose files • Notifying all affected individuals • Determining the severity of the breach • Alerting law enforcement No Insignificant Problem http://epubs.iltanet.org/i/1136335/42 » How to Identify and Report on Compromised Data