The Dark Side of AI and why we need to worry about it?

Artificial Intelligence is being hailed as a brilliant technology that can help businesses increase their operational, predictive, and productive power. Organizations are willing to invest heavily in this disruptive technology that is transforming existing workflows. With AI being omnipresent – (Read – Artificial Intelligence everywhere) – from smartphones to household finances to law and justice systems, it is extremely important that both the public and the companies realize that there could be a potential dark side to artificial intelligence too. There have been innumerable cases of AI bias and violation of the ethics of artificial intelligence in recent times. Forbes describes dark AI as an umbrella term for any evildoing an autonomous system is capable of executing given the right inputs (biased data, unchecked algorithms, etc).

Many movies speculate on how artificial intelligence can escape man’s control to reign supreme, subjecting man to slavery. While this is a bit far-fetched, the actual dangers of AI are much more subtle.

In an article from the Live Mint, Sandipan Deb writes that AI is only as good as the data that is fed into it. “The data is worked on by deep-learning software, which absorbs the data, figures out patterns creates rules to fit the patterns, and keeps tweaking those rules as more data is fed into it.” The data, which is fed by humans, will contain the prejudices that mankind holds, which will ultimately influence the end result that reflects the societal biases like racism, sexism, etc.

Here are a few AI bias examples. In May 2018, a report highlighted that an AI-generated computer algorithm used by a US court for risk assessment was biased against black prisoners. The program asserted that blacks were twice as likely as whites to re-offend in the US. This conclusion was a result of the flawed or skewed training data that it was learning from. While machine learning is often hailed as being impartial and unbiased, the technology will only be as good as the data that has been fed into it. (Read – How AI is learning all our worst impulses)

In 2015, Google came under severe criticism when its photo app tagged two black people as gorillas—perhaps because the algorithm’s training data set did not have pictures of enough black people.

In 2016, Russian scientists ran a global beauty contest to be judged by an AI. Of the 44 winners, only one had dark skin. The algorithm had been trained mostly with photos of white people, and it had equated “fair skin” with “beauty”.

Another example is when an AI is fed the resumes of candidates for a top corporate job, and it chooses a man, because data shows that men have overwhelmingly outnumbered women as CEOs in the past. Going by the data, the AI will decide that a man will make a better CEO than a woman. While the woman would have been pushed back due to gender bias in the past, the computer would not have any idea about this as it is powered by the data it is given.

According to Wikipedia, AI ethics is concerned with the moral behaviour of humans as they design, construct, use and treat artificially intelligent beings, and machine ethics, which is concerned with the moral behaviour of artificial moral agents (AMAs). While AI is being increasingly deployed across a wide variety of domains, from personal digital assistants, email filtering, fraud prevention, voice and facial recognition and content classification to generating news and offering insights into how data centres can save energy, the discrimination that AI could implement should also receive attention. Future of artificial intelligence is indeed very bright but the dark side of artificial intelligence, more than a technical issue, is considered as a social issue of AI and an ethical problem that technologists could find difficult to solve.

Impact of AI in the Telecom Industry

THEN AND NOW

Gone are the days when the telecom industry used to be involved solely in providing basic phone and internet services. The telecom industry trends show that the future of the Telecom Industry is AI-driven. With the bloom of Artificial Intelligence, this particular industry has seen exceptional technological growth.  (Read – Telecoms have unique challenges in adopting AI). The use of AI in the Telecom industry is booming. And it is still growing, with experts expecting it to grow at least 42% by this year.

One move that could transform this sector would be to leverage AI’s ability to engage in active learning while analyzing very large amounts of data collected from their massive customer base.

Where does this data come from?

This data is culled from devices, networks, mobile applications, geolocations, detailed customer profiles, services usage, and billing data. Harnessing this Big data through AI could open up a wide range of uses across management and operations departments.

In a majority of companies, the role of AI in the telecom industry has been limited to chatbots that are automating customer service inquiries, routing customers to the proper agent, and routing prospects with buying intent directly to salespeople.

However, it is also possible to provide better customer experiences, improve operations, and increase revenue through new products and services by gaining actionable insights from data collected. Through the use of AI in the telecom sector, operators can improve network efficiency; lower operating costs, and improve both the quality of service and customer experience.

Tom Anderson, a Principal Technologist at Atis, writes that as operators transition their network architectures with software-defined networking and virtualization technologies that enable automation, AI will leverage these capabilities to self-diagnose, self-heal and self-orchestrate the network.

He says that through the use of algorithms that look for patterns, AI will be able to both detect and predict network anomalies, enabling operators to proactively fix problems before customers are impacted.  This pattern-recognition capability is particularly useful for network security as AI will be able to help identify suspicious activity related to potential security threats, allowing the network to “take-action” in real-time before it impacts network performance.

From a subscriber’s intelligence perspective, AI will allow operators to collect, store, and analyze data from across an operator’s entire customer base to achieve real-time behavioral insights. Their social media, brand coverage, customer sentiments, and other telecom industry benchmarks could also be analyzed to learn what drives customers to the service provider and what drives them to leave.

The information thus gained can be combined with machine learning algorithms to make personalized recommendations based on a user’s behavioral patterns and content preferences. Relevant up-sell and cross-sell offers to the right users at the right time can be made. Data could be analyzed and the call & data package that best suits different types of users can be offered, increasing the sales success rate. AI and machine learning could also be used in detecting and fixing potential issues for market customers of the telecom sector even before they’re apparent to the end-user.

Big data will be essential for operators to achieve better utilization of network resources, allowing the network to adjust services based on user needs, environmental conditions, and business goals resulting in better network optimization.

Cybersecurity and AI

The advancements in computing power and the evolution of paradigms like distributed computing, big data, and cloud computing have brought about an AI revolution. The number of devices that are being connected to the internet is increasing day by day. This also means that the industry is facing more number of cyber-attacks. At the same time, there’s a massive shortage of skilled cyber workers.

With the advancements in AI, many companies have started to use it as a powerful tool against cyber-attacks and cyber-trespassers. AI allows you to automate the detection of threats and combat without the involvement of humans. This can ease the burden on employees, and potentially help identify threats more efficiently than other software-driven approaches. The most primitive form of cyber-attack is spam. Today, machine learning is successfully being used to tackle spam. Google claims that it has a 99 percent accuracy rate in blocking spam. Automatically detecting, analyzing, and defending against attacks enables data deception technology to detect and trick attackers. AI is a machine language-driven, which provides complete error-free cybersecurity services.

That being said, experts say that it is a bit too early for machine learning to be accurate enough in malware detection and prevention. Journalist Marc Ambasna Jones writes that “newer ideas like behavioral analysis and sandboxing powered by ML should be employed in combination with tried and tested techniques, such as firewalls, intrusion detection and prevention, and web and email gateways.” However, companies have also started to put more resources than ever for boosting AI driven technologies.

AI as part of authentication systems

AI is being used for the detection of physical characteristics like fingerprints, retina scans, etc. Thus, biometric logins are much more secure than the password enabled ones. Password protection and authenticity detection systems are vulnerable to attacks and hacks, making the biometric login the better option.

What makes AI vulnerable?

While AI provides solutions to many problems, it can also open up pathways for attacks, especially when it depends on interfaces within and across organizations that inadvertently create opportunities for access by disreputable agents. Nothing stops cybercriminals from using technologies like Machine learning to automatically tailor phishing messages and manipulating data. Deploying AI by the attackers gives them an edge as well, which would enable them to develop automated hacks that can study and learn about the systems they target and identify vulnerabilities.

While AI empowers cyber-security systems, it can also give power to the wrong people. Many believe that AI can truly be a boon to the industry only if it remains in the hands of the right people. Enterprises are faced with challenges of using AI for their profits while balancing the risk of cyber exposure.

Rise of Conversational AI – Going Beyond Simple Chatbots

Up until two years ago, chatbots were hailed as the next big trend. Thousands of chatbots flocked the market as they are relatively easy to build and could be controlled by a predefined flow. However, this trend began to wane when simple chatbots could not meet complex customer demands. It could only work for completing simpler commands like ordering or booking something. Thus, a need for complex chatbots and conversational AI emerged.

Conversational AI was sought out in order to deal with conversations that require a level of comprehension and cognition that goes far beyond the predefined flow of today’s chatbots. Conversational AI is a form of Artificial Intelligence that allows people to communicate with applications, websites, and devices in every day, humanlike natural language via voice, text, touch, or gesture input. Ram Menon, CEO of Awaamo, writes that “these platforms offer more than a natural language interface (NLI): they demonstrate true advancements in combining a variety of emerging technologies — everything from speech synthesis to natural language understanding (NLU) to cognitive and machine learning technologies — and are capable of replacing humans in a variety of tasks.”

Understanding the customer is the key. Using advanced Conversational AI platforms such as Teneo can result not only in an increase in customer satisfaction but in the actionable data that conversational interfaces generate. Such conversational AI-driven chatbots can understand the context and the sentiment behind the conversation. When conversational AI solutions integrate with back-end data and third-party databases, a deeper personalization can take place. It also needs to be capable of creating a detailed analysis of the chat logs in real-time to feedback into the conversation, improve and maintain the system, and deliver actionable insights to the business.

The benefits of using Intelligent Conversational Interface

Intelligent conversational interfaces are the simplest way for businesses to interact with devices, services, customers, suppliers, and employees everywhere. There are lots of companies that provide AI-driven conversational platforms specifically focused on high impact use cases, including IBM’s Watson and KAI.

Analysts predict the rapid and sustained growth of Virtual Digital Assistants in the coming years. This growth underlines the strongly defined benefits that both consumers and enterprises see in conversational AI. The future of chatbots will be dominated by AI-driven conversational tools.

Giving the customers the best experience and analyzing the data garnered can increase the profits of the company.  Businesses can also cut down on costs by using AI-driven chatbots for automating many tasks such as customer service. As cost benefits continue to pile up, the trend will accelerate in 2018.

Artificial Intelligence and the Fourth Industrial Revolution

There are many factors that spike up the production costs of a company. In manufacturing, ongoing maintenance of production line machinery and equipment represents a major expense, having a crucial impact on the bottom line of any asset-reliant production operation especially in this fourth industrial revolution phase. Manufacturing companies are finding it increasingly harder to maintain high levels of quality during the industrial revolution. Bringing out the best product takes time as well as large human resources. But all that is set to change.

Introducing – The Fourth Industrial Revolution

The First Industrial Revolution used water and steam power to mechanize production. The Second Industrial Revolution used electric power to create mass production. The Third Industrial Revolution used electronics and information technology to automate production. Now the Fourth Industrial Revolution is building on the Third and has had a massive impact on the manufacturing sector. The fourth industrial revolution, powered by technology is remolding the industrial sector, helping businesses achieve more profits and more efficiency. The sector is entering its next phase – Industry 4.0 – which is driven by automation, AI, and Internet of things, and cloud computing. The big players are already investing millions in computer intelligence so that they can save time, money, and resources while maximizing their production. The Manufacturer’s Annual Manufacturing Report 2018 found that 92% of senior manufacturing executives believe that ‘Smart Factory’ digital technologies – including Artificial Intelligence – will enable them to increase their productivity levels and empower staff to work smarter.

How is the Manufacturing Sector using Artificial Intelligence?

Through computer vision, machines can be powered to pay attention to the tiniest of details, far beyond a man’s potential. Landing.ai, a startup formed by Silicon Valley veteran Andrew Ng, has developed machine-vision tools to find microscopic defects in products such as circuit boards, using a machine-learning algorithm trained on remarkably small volumes of sample images. If it spots a problem or defect, it sends an immediate alert, an artificial intelligence process known as “automated issue identification.”

Artificial intelligence can also be used to monitor the whole process of manufacturing. Siemens, one of the leading manufacturing companies on the planet, did just that. They embarked on a digitalization strategy of which one of the major goals was Overall Equipment Efficiency. In late 2017, the company announced the latest version of its IoT operating system, MindSphere. Physical machines can be connected to Mindsphere cloud environment, enabling it to build the application that visualizes the various metrics that plant managers need to monitor in 2018. It also gives the resources needed to build an industrial Internet of Things system in a fraction of the time it would take to set up a physical environment.

General Electrics is yet another leader in the manufacturing sector that has adopted the artificial intelligence strategy. In 2015 GE launched its Brilliant Manufacturing Suite for customers, which it had been field testing in its own factories. The system takes a holistic approach of tracking and processing everything in the manufacturing process to find possible issues before they emerge and to detect inefficiencies. GE claims it improved equipment effectiveness at this facility by 18 percent.

It is powered by Predix, their industrial internet of things platform. In the manufacturing space, Predix can use sensors to automatically capture every step of the process and monitor each piece of complex equipment. You can view a short video of how its done here.

Another application of artificial intelligence is the use of generative design. Designers or engineers input design goals into generative design software, along with parameters such as materials, manufacturing methods, and cost constraints. Software explores all the possible permutations of a solution, quickly generating design alternatives. It tests and learns from each iteration what works and what doesn’t. From the many solutions that are put forward, the designers or engineers filter and select the outcomes that best meet their needs. This can lead to major reductions in cost, development time, material consumption, and product weight. Airplane manufacturer Airbus used generative design to reimagine an interior partition for its A320 aircraft and came up with an intricate design that ultimately shaved off 45 percent (30kg) of the weight off the part.

Such applications will affect the future of work. These will bring down labor costs, reduce product defects, shorten unplanned downtimes, improve transition times, and increase production speed. It all comes down to optimal manufacturing performance. If manufacturers do not invest in the long term, ignore the fusion of technologies then their profits would be affected, as prices of products as well as the raw materials would only go up. However, it is not late yet as the field of artificial intelligence is constantly evolving and these applications will be the future of artificial intelligence.

Data Science and how Netflix does it!

Companies of this age invest heavily in data to stay ahead of the competition. Especially the use of data analytics on OTT platform data like those collected from platforms like Amazon’s Prime Video, Hotstar, and Hulu has shown a staggering increase in recent years. This is because of the increasing trend of online streaming entertainment series. Data analysts and data science experts are hired to analyze the vast expanse of data available, an understanding of which will render them capable of making informed and educated choices, optimizing marketing strategies and putting off projects that have a lower rate of success. Once the initial goal of securing customers is achieved, the next step for a successful company is to keep the customers entranced with their work, thereby building the company on their loyalty. For example, recommendation services that are high quality and geared towards recommending new products to customers have the potential to dramatically improve sales per customer and price point per order. Data collected from the customer base can be analyzed to give the customers what they need. Some potential issues can also be solved in the same manner, before they even materialize. Optimizing product locations is another way to provide customer services in a personalized manner.

One of the best examples for analysis of how effective use of big data can result in success is that of Netflix which provides quality content and customer personalization and is presently valued at over 164$ billion, having overtaken Disney’s place as the world’s most valuable media company. Let us dive deep into understanding how Netflix uses data!

Some guidelines to optimize the company’s profits by using data include:

  1. Setting clear goals

Laying out well-defined goals and expectations of the company in order to make the action plan is important as this will give data scientists the opportunity to adopt the right methodology. Netflix is an online streaming channel, with over 130 million users. Their success rates depend on their customer satisfaction, on whether the subscribers like what they are watching. Here Netflix uses predictive analytics to put forth options that the user will find favorable, as it is based on the data collected from the user’s experience. When the channel succeeds to strike a chord with a viewer, then the credibility increases.

  1. Identifying available data

The available data has to be analyzed brought to the right format so that it can be used to bring about the goals of the company. With a large number of subscribers come tremendous amounts of big data on online streaming that Netflix can use. During the onset of the subscription, the customer has to give information about his/her interest in specific genres. They are also asked to rate the movies which they have already seen. Such information is used by Netflix to help them to discover new movies and T.V shows, something that is integral to its success. Data is also collected from events like the customer’s searches, the date the show was watched, the device on which it was watched, when the program was paused, when the program is re-watched, etc. The vast data collected from such events helps Netflix to understand their subscriber’s choices and preferences. This will in turn be used for the user to provide the customer personalization.

3. Adopting the right methodology and being data-driven.

With an in-depth understanding of the data at hand, the right tool has to be chosen so as to use the information most effectively. Netflix uses algorithms for predicting the user’s choice based on his previous ratings. It is said that the recommendation system used by Netflix influences 80% of the content the subscribers watch on Netflix. Recommendation systems are simple algorithms that aim to provide the most relevant and accurate items to the user by filtering useful stuff from a huge pool of information base. Content-based systems recommend items based on a similarity comparison between the content of the items and a user’s profile.

Collaborative Filtering algorithm considers “User Behaviour” for recommending items. Other users’ behavior and preferences over the items are used to recommend items to the new users.

A personalized video ranker orders the entire Netflix collection for each member profile in a personalized way, keeping in mind their interests, habits, and choices. Jenny McCabe, Director of Global Media Relations says “We always use our in-depth knowledge (aka analytics and data) about what our members love to watch to decide what’s available on Netflix….If you keep watching, we’ll keep adding more of what you love.”

An action plan based on the available data would enable the company to make the right decisions. In 2011 Netflix outbid top television channels like HBO and AMC to earn rights for a U.S version of ‘House of Cards.’ At a cost of $4 million to $6 million an episode, this 2-season series cost over $100 million. Such a big decision was made on the data that they already had.

Using methods like clustering analysis, sets with similar attributes are studied. A lot of users watched the David Fincher directed movie The Social Network from beginning to end. The British version of “House of Cards” has also been well watched. Those who watched the British version of “House of Cards” also watched Kevin Spacey films and/or films directed by David Fincher. Such factors gave them the confidence to make the $100 million investment, which turned profitable in the end. Here, using association mining of customer information with similar behavior is targeted to make a decision that satisfies that particular set.

  1. Testing regularly

Data-driven models should be constantly checked as demographics have a way of changing gradually. The algorithms that Netflix uses are constantly revised in order to achieve maximum optimization. Bill Franks, Chief Analytics Officer, International Institute for Analytics says that “ I can say that no changes in Netflix products are not tested and validated and we do not just test to test. If we do not believe it will not improve, it will not be tested. We have 300 major tests of products and dozens of variations within”

Data science practices should be implemented wisely. Netflix has successfully shown us that machine learning can be used to convert the user’s cravings into the company’s business goal. Quantitative data is always a good basis on which better and cost-effective decisions can be taken. Data can predict whether certain innovations or experimental projects can take off.

While discussing analytics, Netflix co-founder Mitch Lowe says “He [Reed Hastings] taught me how to use analytics to make decisions. I always thought you needed a clear answer before you made a decision and the thing that he taught me was [that] you’ve got to use analytics directionally…and never worry whether they are 100% sure. Just try to get them to point you in the right direction. ”

Intelligent use of data can reap benefits, but it should be done responsibly. Data should be protected from others and used carefully. Data science should be seen as a solution to solving problems as well as a way to greater rewards; it should be given due importance.