Artificial intelligence (AI) has been making great strides in recent years sometimes even coming close to being human-like. Now, in a new paper published in Nature magazine, IBM describes a system that can debate with humans and even sometimes win.

“Here we present Project Debater, an autonomous debating system that can engage in a competitive debate with humans,” write the authors. And the system is nothing short of extraordinary.

In tests of Project Debater, the AI was given only 15 minutes to research topics and prepare for debates. Each time, it proceeded to form an opening statement and even layer counterarguments. 

For the most part, the humans won the debate but in one instance it was able to change the stance of nine people. Not bad!

“Project Debater is a crucial step in the development of argument technology and in working with arguments as local phenomena. Its successes offer a tantalizing glimpse of how an AI system could work with the web of arguments that humans interpret with such apparent ease,” Chris Reed writes in a critique of the new project published in Nature magazine.

“Given the wildfires of fake news, the polarization of public opinion and the ubiquity of lazy reasoning, that ease belies an urgent need for humans to be supported in creating, processing, navigating and sharing complex arguments — support that AI might be able to supply.”

In other words, this new AI is not here to replace humans but rather to support them in building better arguments and reasoning with more nuance. If this subject interests you, Scientific American has done a great podcast episode with the research’s lead Noam Slonim which tackles amongst other things whether the AI actually understands the arguments it presents and what that means for the future of debating.

The post IBM Developed an AI System That Engages in Debates with Humans and Convinces Some appeared first on Artificial Intelligence.

Eleanor Brodie, data science manager at LexisNexis Risk Solutions tells Insurance Times about the biggest misconceptions with data in the industry and why tapping into it could be an asset

What are the biggest misconceptions you encounter about data for insurance risk assessment?

A big misconception in the industry is that big data and analytics driven products will replace human capital. Data-driven products allow insurance companies to streamline their existing processes and are meant to complement their existing workflow. Human judgment and expertise will always be required to set the foundation to ensure risk is accurately priced and aligns with each company’s business strategy. However, data-driven insights can assist with the decision process.

Another misunderstanding about data science is there is a simple formula where all data is put into a magic funnel that draws out the desired outcome. However, before any predictive model can be built or nugget of insight can be gleaned, data generally needs to be enriched, filtered, and structured and this process relies heavily on the quality of the initial data sources and how they are modelled.

How much has changed for the data science team at LexisNexis Risk Solutions in the past year both operationally and in terms of focus as a direct consequence of the pandemic?

Aside from the adjustment to homeworking which has entailed lots of video calls with customers and colleagues, the past year has certainly fuelled the appetite for more and more data, attributes, and scores. Insurance providers are hungrier now than they have ever been to evaluate new data driven solutions to better understand and segment their customers.

A significant challenge for insurance providers is any sudden change in consumer behaviour. We all do things we do not even think about as part of our daily routines. Shopping for insurance, driving to work and going to the supermarket are just a few activities that just happened until the pandemic took hold. With very little notice, insurance providers needed to augment and expedite their ability to service their prospects and customers virtually.

This includes prefill solutions, data driven underwriting, pricing applications, and contactless claims processing. Fortunately, we have been developing these solutions for years, so we have been in a good place to help insurance providers interact effectively with their customers and understand how risks have changed as a direct consequence of the pandemic.

Data is often called ‘the new gold’, but what process does the business go through to assess the market opportunity and bring datasets to insurers and brokers to ensure they are satisfying an insurance market need?

We start by building an analytical prototype to validate an idea from our colleagues in the home, motor, or commercial teams. These ideas come from the constant dialogue we have with customers on the pain points they need solved. Our new Covid-19 attributes are a good example of this – we could see the value the market could gain from understanding changes in motor policy behaviour during the first lockdown versus changes outside of that time.

Once we have proved the concept, we start the product development work which may leverage millions of data points. We then create the final specs for technology to implement. As we near implementation, we develop the attributes or inputs into the solution and our audit team works with technology to ensure the final product performs as expected.

As the product is being developed, we give our customers the opportunity to test the solution on their own data. This might be through actionable insight studies or we may perform retro validations tests through our batch team.

Closer to launch, we look at any required regulatory documents on the solution inputs, outputs, and overall performance. Then following launch we monitor the attributes and scores to ensure they continue to perform as expected.

It’s a well-oiled process and can be hugely rewarding when you find a new data attribute that you know can solve a pain point for the market.

Given the increasing demand for data scientists across many industries, what makes working with insurance data attractive to candidates and how do you attract and nurture this talent?

Competition for data scientists is indeed hot and while there are more colleges delivering great talent into the marketplace, I think demand will exceed supply for some time yet.

I might be biased but we do have a lot more to offer good data scientists – we have what they dream of – good quality data and lots of it. Our data scientists work with literally hundreds of millions and often billions of records to solve our customers’ problems. Where other companies are limited in the breadth and depth of their data and the ability to pull it all together in a commercially viable application, this is what we do every day and that is exciting to a candidate.

Our Data Science Rotational Program (DSRP) sees graduates from a wide array of disciplines, including but not limited to mathematics, statistics, computer science, data science, physics, financial math, actuarial science, and engineering join LexisNexis for a two-year cycle through four different teams.

This experience provides a robust hands-on journey from data access, data analysis, model building to model implementation. Right now, we have six DSRP team members in this programme globally and we typically hire three new positions each year.

How can brokers and insurers put themselves in the strongest position to use data as part of the customer journey – from quoting to claim as we head into 2021?

Look for data enrichment solutions that can be pulled into existing processes seamlessly and efficiently. Calling out for data from multiple data sources is inefficient and can lead to a poor customer experience so look for platforms solutions that offer the widest choice of data. You may not want that whole choice today but you do want the flexibility to expand in the future.

Consider joining data sharing initiatives which will help you gain a view of the industry’s experience of an individual, vehicle or location. Shared motor policy history data has already established links between policy behaviour and claims losses. The next development will focus on shared claims data.

Also conquer the challenge of creating a single customer view. Many insurance providers have gone through merger and acquisition activity over the past few years which has made customer database management a real headache. It means you can end up with multiple records for the same customer. When you are able to pull those disparate records together through linking and matching technology to create one consolidated view, you open up cross sell and upsell opportunities based on a much clearer understanding of the customer. It also helps you gain the most value from data enrichment solutions.

Finally, for insurance providers to truly maximise and leverage the opportunities new data sources will bring through data enrichment at point of quote and renewal, it’s worth going back to their initial data sources. 

This can start with refreshing the initial data model. Refreshing the data will identify behaviours used in rating the risk, how it’s changed and how it should be adapted for the current market. Without this crucial first step, adding in new data could be duplicating effort and capturing behaviours that existing models may already capture.

It is also worth viewing new data sources as possible replacements for existing data sets rather than an add on to the current data used. Insurance providers need to look at the incremental benefits a new data source will bring.

As the industry continues to rise to the challenge of pricing in a highly competitive market, maximising the opportunities of its data models to create more intelligible insights and a strategic advantage over competitors is vital.

The post Q&A: Big data analytics driven products replacing humans a ‘misconception’ – LexisNexis Risk Solutions appeared first on Artificial Intelligence.

By learning about R and Data science, humans are provided with ample of opportunities in the world of data.

The education about data science is not enough. The more we read and learn about data science, the more we become fascinated about the intricate learning data science has to offer. Since data science is the new hype and will continue to remain so in the future, here are top 10 free online books that are coherent and comprehensive to understand R/Data science.

• 1. Advanced R by Hadley Wickham- Aiming at the intermediate and advanced users, the book talks about the fundamentals of R and the data types, and solving wide range of programs using functional programming. This book is a must go if one has to make the R code faster and efficient.

• 2. Introduction to Data Science by Rafael Irizarry- Introducing the concepts and skills for solving data analysis challenges, this book covers the concepts of probability, statistical interference, linear regression and machine learning. Moreover, this book assist in developing skills pertaining to R programming, data wrangling with dplyr, data visualization with ggplot2 and algorithm building with caret amongst others.

• 3. Cookbook for R by Winston Chang- Being a fantastic resource for getting started about plotting with ggplot and more, this book offers answers to lots of coding questions, which arise while making publication quality graphics with R.

• 4. Data Visualization: A practical introduction by Kieran Healy- Offering a hands-on introduction about visualization data using R and Wickham’s ggplot, this book assist in building the visualisations for data science piece by piece, from simple scatter plots to more complex graphics.

• 5. Exploratory Data Analysis with R by Roger D Peng- Based on the courses from John Hopkins Data Science Specialization, this book covers the basics in exploratory analysis, and topics needed for analyzing and visualising high-dimensional or multi-dimensional data like Hierarchial clustering, K-means clustering, and dimensionality reduction techniques-SVD and PCA.

• 6. Text Mining with R: A Tidy approach by Julia Silge and David Robinson- Being a great introductory book to learn about mining text data with R, this book helps in practicing the principles in text datasets. Moreover, using R and tideverse as examples to explore literature, news, social media data, this book is a must go for learning about text and data analysis, specifically for those who are interested in analysing the social media data.

• 7. An Introduction to Statistical and Data Sciences via R by Chester Ismay and Albert Y.Kim- Covering the basics of statistics for data science using R, this book helps in learning about exploring data, basics of statistics for data science and creating data stories using R.

• 8. Introduction to Empirical Bayes: Examples from Baseball Statistics by David Robinson- Introducing the empirical Bayesian approach for estimating credible intervals, A/B testing and mixture models with R code examples, this book illustrates statistical method for estimating click-through rates on ads, and success of experiments amongst others. This book is a must go if one wants to learn about data science and statistics for data science.

• 9. Data Analysis for the Life Sciences with R by Rafael A Irizarry and Michael I Love – Primarily focusing on high throughput data from genomics, the book helps the reader to solve problems with R code and assist in gaining better intuition behind the math theory. The methods described in this book are best suited for modern data science in any domain.

• 10. Modern Data Science for Modern Biology by Susan Holmes- With only 13 chapters, this book is a comprehensive guide for beginners to learn about R code, theory, and great visualization with ggplot 2. This book also covers various aspects of statistics for data science including, Mixture models, clustering, testing, dimensionality reduction techniques like PCA and SVD.

The post TOP 10 FREE ONLINE BOOKS TO LEARN R-CODE AND DATA SCIENCE appeared first on Artificial Intelligence.

Since the dawn of Bronze age civilizations more than 5000 years ago, humans have been creating norms of societal governance. The process continues with many imperfections. Off late, Artificial Intelligence (AI) is increasing its influence in decision making processes in the lives of humans and expectations are whether AI will follow similar or better norms. Principles that govern the behaviour of responsible AI systems are being established.

Principles

Fair

All AI systems should be fair in dealing with people and be inclusive in coverage. In particular, they should not show any bias in working. Historically, humans have used at least 2 major criteria for unfair treatment, i.e. gender and caste/race/ethnicity. 

Amazon tried to develop an algorithm for recruitment. However, it started showing less tendency to select female candidates. Even after removing gender specific indicators, females were still discriminated against. The project had to be abandoned.

Compas, a risk-assessment tool developed by a privately held company and used by the Wisconsin Department of Corrections predicted that people of colour had higher tendency of repeat offences than they actually do. California has decided not to use face recognition technology for law enforcement. A study by Stanford researchers in 2020 found that voice recognition software of Amazon, Apple, Google, IBM, and Microsoft have higher error rates when working on voice of black people.

Transparent and Accountable

Unlike traditional software, it is hard to predict the outcome of AI algorithms as they dynamically change with training. This makes them less transparent and this “Black box” nature of AI makes it very difficult to find the source of error in case of a wrong prediction. This also makes it makes difficult to pinpoint accountability. Neural networks are the underlying technology for many face, voice, character etc recognition systems. Unfortunately, it is more difficult to trace problems in neural networks especially deep ones (with many layers) than in other AI algorithms e.g decision trees etc. And new variants of neural networks e.g. GANs (Generational Adversarial Networks), Spiking Neural Networks etc continue to gain popularity.

Reliable and safe

Security and reliability of AI systems has certain peculiar dimensions e.g. unpredictability. Facebook in collaboration with Georgia Institute of Technology created bots that could negotiate but they also learnt how to lie. This was not intended during programming. Another issue is slow rise of Artificial General Intelligence (AGI) or Broad AI or Strong AI that aims to create systems that genuinely simulate human reasoning and generalize across a broad range of circumstances. These algorithms will be able to do transfer learning, so an algorithm that learns to play Chess will also be to able learn how to play Go. This will vastly increase the context in which a machine can operate and this cannot be predicted in advance.

Unpredictability reduces reliability and safety of the systems.

Problem sources

Models and features

The power of AI algorithms is based on the models and features and the weightages of the features that are used while creating models. The AI in use currently is also Narrow AI and it will not work if the context changes. For example, a system designed to scrutinize applications for medical insurance policies may discriminate against people with diseases if used to vet applications for car insurance since the features and their weightages are not appropriate for the latter case. Hence models or features framed without fairness in mind can induce biases.

Data

The biggest source of biases in AI systems is data as biases may be inherent in the data, either explicitly or subconsciously. This can happen if data is not uniformly sampled or carries implicit historical or societal biases. In credit risk, data of customers who defaulted less as they were supported by tax benefits will give incorrect results when used for scenarios where tax benefits are not there. MIT researchers found that facial analysis technologies had higher error rates for minorities and particularly minority women, potentially due to unrepresentative training data. The reason for failure for Amazon recruitment software was that it was trained on 10 years of data where resumes of male candidates outnumbered that of females. It also focussed on words e.g. “executed”, “captured” etc that are more commonly used by males.

Other issues

The rise of AI poses additional challenges not found in traditional systems.

Driverless Vehicles

The driverless vehicles will start plying on the roads in a decade or so. Any accident will raise the question of civil and criminal liability. In 2018 a pedestrian died when she was hit by Uber test car despite a human driver sitting inside the car. A vehicle may be programmed to save either the passengers or pedestrians. Potential accused could be vehicle manufacturer, vehicle operator or even the government. This will also change the underwriting models. Liability issues will also come as companies allow operation decisions to be more data driven as now programmers will appear to be the sole accused. 

Weapons

Countries e.g. US, Russia, Korea etc plan to use AI in weapons e.g. drones or robots etc. Currently the machines do not have emotions and this raises the concern if an autonomous machine goes on killing spree. In 2018, Google had to stop engagement with US government over its Maven military program due to public outcry.

Safeguards

Guidelines

The concerns over ethics in AI have resulted in many organizations formulating guidelines governing the use of AI e.g. European Commission’s, “Ethics Guidelines for Trustworthy Artificial Intelligence”,  US government’s “Roadmap for AI Policy”, IEEE’s P7000 standards projects etc. These contain the general principles of ethics and responsibility that AI systems should follow.

Software

Many companies have created frameworks, software, guidelines etc that can help to create Responsible AI e.g. IBM, Google, Microsoft, PWC, Amazon, Pega, Arthur, H2O etc. Their software help to explain model’s “Black box” behaviour and hence bring transparency, assess fairness of the systems, mitigate bias against any identity based groups, keep the data secure etc by constant monitoring.

Companies

Within companies, Responsible AI can be facilitated by imposing standards through overseeing groups, creating diversity in teams and cascading the message to individuals. There should be conscious efforts to reduce biases in data.

Future

In the next two decades, machines will become more autonomous in decision making processes and human will slowly cede control of their own lives. Establishment of Responsible AI will reduce biases and increase acceptance of AI. This will help in creating a more fair and equitable society. An unchecked growth of AI will not only make humans less tolerant to AI but also to each other.

The post Artificial Intelligence Must Be More Responsible Than Humans appeared first on Artificial Intelligence.

Giving robots, who currently rely on vision and touch to move around, power to hear sounds and predict the physical properties of objects around them can be a game changer, say researchers including two of Indian origin.

People rarely use just one sense to understand the world, but robots usually only rely on vision and, increasingly, touch.

The researchers from Carnegie Mellon University (CMU) now say that robot perception could improve markedly by adding another sense: hearing.

“A lot of preliminary work in other fields indicated that sound could be useful, but it wasn’t clear how useful it would be in robotics,” said Lerrel Pinto, who recently earned his PhD in robotics at CMU and will join the faculty of New York University this fall.

He and his colleagues found the performance rate was quite high, with robots that used sound successfully classifying objects 76 per cent of the time.

The team at CMU’s Robotics Institute found that sounds could help a robot differentiate between objects, such as a metal screwdriver and a metal wrench.

Hearing also could help robots determine what type of action caused a sound and help them use sounds to predict the physical properties of new objects.

Pinto said that the results were so encouraging that it might prove useful to equip future robots with instrumented canes, enabling them to tap on objects they want to identify.

The researchers presented their findings last month during the virtual Robotics Science and Systems conference.

Other team members included Abhinav Gupta, associate professor of robotics, and Dhiraj Gandhi, a former master’s student who is now a research scientist at Facebook Artificial Intelligence Research’s Pittsburgh lab.

To perform their study, the researchers created a large dataset, simultaneously recording video and audio of 60 common objects — such as toy blocks, hand tools, shoes, apples and tennis balls — as they slid or rolled around a tray and crashed into its sides.

They have since released this dataset, cataloging 15,000 interactions, for use by other researchers.

The team captured these interactions using an experimental apparatus they called Tilt-Bot — a square tray attached to the arm of a Sawyer robot.

It was an efficient way to build a large dataset; they could place an object in the tray and let Sawyer spend a few hours moving the tray in random directions with varying levels of tilt as cameras and microphones recorded each action.

They also collected some data beyond the tray, using Sawyer to push objects on a surface.

Pinto said the usefulness of sound for robots was therefore not surprising, though he and the others were surprised at just how useful it proved to be.

“I think what was really exciting was that when it failed, it would fail on things you expect it to fail on,” he said.

For instance, a robot couldn’t use sound to tell the difference between a red block or a green block.

“But if it was a different object, such as a block versus a cup, it could figure that out.”

The post Robots who can hear work more like humans: Researchers appeared first on Artificial Intelligence.

Artificial super-intelligence (ASI) is a software-based system with intellectual powers beyond those of humans across an almost comprehensive range of categories and fields of endeavor.

The reality is that AI has been with here for a long time now, ever since computers were able to make decisions based on inputs and conditions. When we see a threatening Artificial Intelligence system in the movies, it’s the malevolence of the system, coupled with the power of some machine that scares people.

However, it still behaves in fundamentally human ways.

The kind of AI that prevails today can be described as an Artificial Functional Intelligence (AFI). These systems are programmed to perform a specific role and to do so as well or better than a human. They have also become more successful at this in a short period which no one has ever predicted. For example, beating human opponents in complex games like Go and StarCraft II which knowledgeable people thought wouldn’t happen for years, if not decades.

However, Alpha Go might beat every single human Go player handily from now until the heat death of the Universe, but when it is asked for the current weather conditions there the machine lacks the intelligence of even single-celled organisms that respond to changes in temperature.

Moreover, the prospect of limitless expansion of technology granted by the development of Artificial Intelligence is certainly an inviting one. While investment and interest in the field only grow by every passing year, one can only imagine what we might have to come.

Dreams of technological utopias granted by super-intelligent computers are contrasted with those of an AI lead dystopia, and with many top researchers believing the world will see the arrival of AGI within the century, it is down to the actions people take now to influence which future they might see. While some believe that only Luddites worry about the power AI could one-day hold over humanity, the reality is that most tops AI academics carry a similar concern for its more grim potential.

Its high time people must understand that no one is going to get a second attempt at Powerful AI. Unlike other groundbreaking developments for humanity, if it goes wrong there is no opportunity to try again and learn from the mistakes. So what can we do to ensure we get it right the first time?

The trick to securing the ideal Artificial Intelligence utopia is ensuring that their goals do not become misaligned with that of humans; AI would not become “evil” in the same sense that much fear, the real issue is it making sure it could understand our intentions and goals. AI is remarkably good at doing what humans tell it, but when given free rein, it will often achieve the goal humans set in a way they never expected. Without proper preparation, a well-intended instruction could lead to catastrophic events, perhaps due to an unforeseen side effect, or in a more extreme example, the AI could even see humans as a threat to fully completing the task set.

The potential benefits of super-intelligent AI are so limitless that there is no question in the continued development towards it. However, to prevent AGI from being a threat to humanity, people need to invest in AI safety research. In this race, one must learn how to effectively control a powerful AI before its creations.

The issue of ethics in AI, super-intelligent or otherwise, is being addressed to a certain extent, evidenced by the development of ethical advisory boards and executive positions to manage the matter directly. DeepMind has such a department in place, and international oversight organizations such as the IEEE have also created specific standards intended for managing the coexistence of highly advanced AI systems and the human beings who program them. But as AI draws ever closer to the point where super-intelligence is commonplace and ever more organizations adopt existing AI platforms, ethics must be top of mind for all major stakeholders in companies hoping to get the most out of the technology

The post IS DYSTOPIAN FUTURE INEVITABLE WITH UNPRECEDENTED ADVANCEMENTS IN AI? appeared first on Artificial Intelligence.

Malware, phishing and ransomware are constantly keeping security teams on their toes. But there is one risk to data security that cannot be stopped by cybersecurity software: human error.

Unlike malicious threat actors, human error doesn’t come and go as trends in the cyber landscape change. It is true of data breaches throughout history: indeed, a CybSafe study found that human error caused 90% of cyber data breaches in the UK during 2019. For organisations looking to protect intellectual property, or shield customer data, human error is the most dangerous threat of all. And yet, it appears to have no foolproof solution, or remedy available to buy from a cybersecurity vendor.

Humans are naturally prone to making mistakes. Such errors are increasingly impactful in the workplace, but human error in the realm of cybersecurity can have particularly devastating and long-lasting effects. As the digital world becomes more complex, it becomes much tougher to navigate – and thus, more unfair to blame humans for the errors they make. Employees should be given as much help and support as possible.

But employees are not often provided with the appropriate security solutions, so they resort to well-intentioned workarounds in order keep pace and get the job done. As data continues to flow faster and more freely than ever before, it becomes more tempting to just upload that document from your personal laptop, or click on that link, or share that info to your personal email.

Take, for instance, one of the most common security problems: phishing emails. An employee might follow instructions in a phishing email not only because it looks authentic, but that it conveys some urgency (usually from a manager or someone else of importance). Employee training can help reduce the likelihood of error, but solving the technological shortcoming is more effective: if a phishing email is blocked from delivery in the first place, we can help mitigate the human error factor. 

This is where artificial intelligence (AI) can be a game-changer. We already use AI to simplify our home lives, using it to perform a variety of tasks, from turning on lights, to playing our favourite music. But if AI solutions are deployed in the workplace, we can help address the biggest elephant in the IT room: data security.

Data security is a major area of concern, and it’s likely the leading cause for lost hours – and lost sleep – for security and IT professionals. According to a recent survey of over 500 IT professionals in the financial services industry, a whopping 94% said that they lack confidence in the ability of employees, consultants, and partners to safeguard customer data. And because cybersecurity is a complex domain – with many unknowns and moving parts – the rigid, conventional solutions can’t be effective. However, AI solutions can learn, adapt, and dynamically react to an organisation’s cybersecurity needs.

Not to worry, though – this is a far cry from the sensationalistic sci-fi scenes of a robot takeover. Yes, AI can solve complex problems with a level of consistency and speed that’s unmatched by human intelligence. But it can’t replace human intelligence where it’s needed most: we must choose the right problems to solve. Once we identify this, we can start collecting the right data, designing the right solution, and creating the right processes for our AI solutions to adapt, learn from feedback, and produce results.

AI can also provide employees with more time to tackle the impactful tasks of the business. Consider the case of phishing emails again: Even if the employee properly deletes that email, they’ve still spent valuable time in security training sessions, and in evaluating that email for potential threats. With an AI-based solution that detects phishing emails before they’re delivered, the employee’s time and efforts can be much better spent.

If AI can eliminate human error from businesses, will humans still be valuable to organisations? Of course – and their value will be vastly increased, too. AI helps to augment human intelligence, by taking on the low-level tasks and conducting the meticulous checks that the current cybersecurity landscape requires. This gifts humans more time to achieve their full potential in their role, from solving problems using critical thinking, to developing creative ideas from insights. For organisations, this is the ultimate in added value – not only can human error be reduced, but the true potential of teams is finally unlocked.

The post Empower humans by deploying AI for cybersecurity appeared first on Artificial Intelligence.

Artificial intelligence and machine learning models can work spectacularly — until they don’t. Then they tend to fail spectacularly. That’s the lesson drawn from the COVID-19 crisis, as reported in MIT Technology Review. Sudden, dramatic shifts in consumer and B2B buying behavior are, as author Will Douglas Heaven put it, “causing hiccups for the algorithms that run behind the scenes in inventory management, fraud detection, marketing, and more. Machine-learning models trained on normal human behavior are now finding that normal has changed, and some are no longer working as they should.”

Machine-learning models “are designed to respond to changes,” he continues. “But most are also fragile; they perform badly when input data differs too much from the data they were trained on. It is a mistake to assume you can set up an AI system and walk away.” 

It’s evident, then, that we may be some ways off from completely self-managing systems, if ever. If this current situation tells us anything, it’s that human insights will always be an essential part of the AI and machine learning equation. 

In recent months, I had been exploring the potential range of AI and machine learning with industry leaders, and what role humans need to play. Much of what I heard foreshadowed the COVID upheaval. “There is always the risk that the AI system makes bad assumptions, reducing performance or availability of the data,” says Jason Phippen, head of global product and solutions marketing at SUSE. “It is also possible that data derived from bad correlations and learning are used to make incorrect business or treatment decisions.  An even worse case would clearly be where the system is allowed to run free and it moves data to cold or cool storage that causes loss of life or limb.”   

AI and machine learning simply can’t be dropped into an existing infrastructure or set of processes. Chris Bergh, CEO of DataKitchen, cautions that existing systems need to be adapted and adjusted. “In traditional architecture, an AI and machine learning system consumes data environments to fulfill the data needs,” he says. “We need a slight change to that architecture by letting AI manage the data environment. This transition must be done smoothly in order to prevent catastrophic failures in the existing systems as well as to implement robust systems.”

AI and machine learning systems “being developed to manage data environments must be considered as mission-critical systems, and the development must be carried out very carefully,” Bergh continues. “Since data is the driving force of present-day business decisions, data environments will be the heart of the business. Therefore, even a slight failure in data management will incur a significant cost to the business by loss of operational time, other resources and user trust.”

Bergh also points to the “knowledge gaps of data professionals and AI and machine learning experts in the areas of AI and machine learning and data management, respectively.”   

The bottom line is that skilled humans will always be key to managing the flow and assuring the quality and timeliness of data being fed into AI and machine learning systems. The mechanics of data management will be autonomous, but the context of the data needs human involvement. “We can look at examples like self-driving cars and data center energy optimization using DeepMind at Google and be fairly confident that there will eventually be some parallel opportunities in database management,” says Erik Brown, a senior director in the technology practice of West Monroe Partners, a business/technology advisory firm. “However, fully autonomous databases are likely a stretch in the near future; human involvement should become more strategic and focused in areas where humans are best equipped to spend their time.” 

Fully autonomous data environments “will likely take many years to achieve,” agrees Jeremy Wortz, a senior architect in West Monroe’s technology practice. “Machine learning is far from solving complex wide problems. However, an approach that develops narrow and deep use cases will make a difference over time and will start the journey of a self-managing system. Most organizations can take this approach but will need to ensure they have a way to enumerate the narrow use cases, with the right tech and talent to realize these use cases.”

The more organizations depend on AI, the more humans will need to step up and oversee the data that is moving into these systems, as well as the insights that are being produced. Eighty percent or more of the effort in AI and machine learning “is often data sourcing, translation, validation and preparation for complex models,” says Brown. “As these models are informing more critical business use cases — fraud detection, patient lifecycle management — there will continue to be more demands on the stewards of that data.”

Few data environments outside of the Googles and Amazons of the world are truly ready, Brown says. “This is a huge opportunity for growth in most industries. The data is there, but collaborative, cross-functional organizational structures and flexible data pipelines aren’t ready to harness it effectively.”

One does not have to be a degreed data scientist to manage AI systems — what is needed is an interest in learning and leveraging new techniques. “AI-powered technology is fueling the citizen data scientist trend, which is a game-changer,” says Alan Porter, director of product marketing at Nuxeo. “In the past, these roles have required deep technical knowledge and coding skills. But with advances in technology — many of the tools and systems do the heavy technical lifting for you. It’s not as critical for people to fill these positions to have technical knowledge, instead organizations are looking for people who are more analytical with specific business expertise.” 

While people with technical and coding skills will still play a critical role within organizations, Porter continues, “a big piece of the puzzle is now having analysts with specific business knowledge so they can interpret the information being gathered and understand how it fits into the big picture. Analysts also have to be good at communicating their findings to stakeholders outside the analytics team in order to effect change.”   

In his MIT piece, Heaven concludes that “with everything connected, the impact of a pandemic has been felt far and wide, touching mechanisms that in more typical times remain hidden. If we are looking for a silver lining, then now is a time to take stock of those newly exposed systems and ask how they might be designed better, made more resilient. If machines are to be trusted, we need to watch over them.” Indeed. 

The post No matter how sophisticated, artificial intelligence systems still need human oversight appeared first on Artificial Intelligence.

Leveraging AI’s full potential doesn’t mean developing a pilot project in a vacuum with a handful of experts – which, ironically, is often called accelerator project. Companies need a tangible idea as to how artificial intelligence can benefit them in their day-to-day operations.

For this to happen, one has to understand how these new AI ‘colleagues’ work and what they need to successfully do their jobs.

An example for why this understanding is so crucial is lead management in sales. Instead of sales team wasting their time on someone who will never buy anything, AI is supposed to determine which leads are promising and at what moment salespeople can make their move to close the contract. CEOs are usually very taken with that idea, sales staff not so much.

Experienced salespeople know that it’s not that easy. It’s not only the hard facts like name, address, industry or phone number that are important. Human sales people consider many different factors, such as relationships, past conversations, customer satisfaction, experience with products, the current market situation, and more.

Make no mistake: if the data are available in a set framework, AI will also leverage them, searching for patterns, calculating behavior scores and match scores, and finally indicating if the lead is promising or not. They can make sense of the data, but they will never see more than them.

The real challenge with AI are therefore the data. Without data, artificial intelligence solutions cannot learn. Data have to be collected and clearly structured to be usable in sales and service.

Without big data no AI

Without enough data to draw conclusions from, all decisions that AI makes will be unreliable at best. Meaning that in our example, there’s no AI without CRM. That’s not really new, I know. However, CRM systems now have to be interconnected with numerous touchpoints (personal conversations, ERP, online shops, customer portal, website and others) to aggregate reliable customer data. Best case: all of this happens automatically. Entrusting a human with this task makes collecting data laborious, inconsistent and faulty.

To profit from AI, companies need to understand where it makes sense to implement it and how they should train it. There’s one problem, however: the ‘thought patterns’ of AI are often so complex and take so many different information and patterns into consideration that one can’t understand why and how it made a decision.

In conclusion, AI is not a universal remedy. It’s based on things we already know. Its recommendations and decisions are more error-prone than many would like them to be. Right now, AI has more of a supporting role than an autonomous one. They can help us in our daily routine, take care of monotonous tasks, and let others make the important decisions.

However, we shouldn’t underestimate AI either. In the future, it will gain importance as it grows more autonomous each day. Artificial intelligence often reaches its limits when interacting with humans. When interacting with other AI solutions in clearly defined frameworks, it can often already make the right decisions today.

The post Artificial Intelligence (AI) Is Nothing Without Humans appeared first on Artificial Intelligence.

Artificial intelligence (AI) is ushering in a paradigm shift in the conservative insurance industry, as more market players introduce new sets of time-saving AI-powered platforms to enhance business efficiency.

It has been years since AI started appearing in almost all industries due to its usefulness in a wide range of applications and human-like versatility.

Most globally leading companies have gone all-out to incorporate the big data-driven technology into their platforms, in a desperate bid to take the lead in the promising growth area.

Despite the global AI sensation, local insurers have so far remained less agile in jumping on the bandwagon. But starting this year, they began to utilize the technology in their own ways by developing automated, data-driven AI software.

On the development front, the Korea Insurance Development Institute (KIDI) recently developed an AI-driven car repair estimation software program. This will streamline insurance payment processes, as the platform _ called AOS Alpha _ can calculate the cost of repairs by analyzing images of a damaged car.

The institute developed the AI platform by analyzing 1 million images via deep learning. It said AOS Alpha can estimate repair costs for 195 cars _ ranging from sedans to SUV. The figure will increase down the road, as the platform increases its datasets. The accuracy of the platform will also improve against the same backdrop.

The nation’s 12 non-life insurers, which operate car insurance, plan to bring in the service for their businesses. KIDI said they are training staffs before introducing the platform.

“The introduction of the platform in the local insurance industry will help improve insurance culture by reducing distrust and conflict between interested parties, as AOS Alpha is expected to set a standard in offering car repair costs,” said Park Jin-ho, research head at KIDI’s auto technology laboratory.

Staring from mid-January, Hanwha Life Insurance, one of the so-called big three life insurers here, also introduced an AI-driven insurance claim review system. This was the first time in the industry that a cloud-based platform was introduced for real-time insurance claim analysis here.

According to the company, the system can automatically make decisions on whether to provide insurance payments to its users. The company has racked up 11 million insurance claim datasets over the past three years, ever increasing the accuracy of the system.

For now, the system processes 25 percent of insurance claims, but the company plans to increase the figure to 50 percent in the near future. The Hanwha affiliate expects the platform to reduce fixed costs of more than 10 billion won ($8.16 million) for the next five years.

Kyobo Life Insurance and Samsung Fire & Marine Insurance are also joining the AI drive by utilizing the technology in their insurance contract review process.

In October 2019, Kyobo Life developed BARO insurance contract review system. According to the company, BARO is capable of making human-like decisions in whether to approve an insurance contract. The AI system analyzes whether an applicant meets the standards of its insurance product.

The Samsung affiliate is also operating the AI-powered insurance contract review system. The software can automatically approve or reject users’ applications for the firm’s health and cancer insurance products by conducting computerized screening procedures without human labor, the company said.

The post AI taking center stage in insurance industry appeared first on Artificial Intelligence.