Bucket testing: marketing or social engineering?

30

September

2018

No ratings yet.

As an individual user of online platforms you may not notice on your own, but companies are conducting experiments over and over again to get a better understanding of … You! This blogpost will introduce one of the methods how companies experiment and acquire more information about the effectivity of their own platform and how they should approach you.

Bucket testing, also known as A/B testing or split testing, is a method of comparing two or more versions of a webpage, app or ad with each other in order to determine which one performs better. To give an example, imagine two variants of an ad. The former version is a video ad and the latter version a simple banner. By exposing users on a random basis to one of the two variants, companies can use statistical models to determine which ad performs better (e.g. measure the number of clicks). Facebook offers its platform for such practices, allowing companies to acquire data in the best way to reach their audience.

From a marketing perspective, bucket testing is a strong tool to gather information about the behaviour of an individual user through hypothesis testing. A user may be for instance more inclined to click on a certain object depending on the size, the colours or even its position on the webpage. For companies, this would mean that effective use of bucket testing leads to more exposure, as the ads are specifically tailored so that the audience is more likely to interact with it.

While bucket testing is not new phenomenon, it does raise some issues. A platform may become so powerful after thousands or millions of tests, that it can accurately leverage on people’s personal vulnerabilities, and influence an individual to take an action that may or may not be in his or her best interest. The possibility exists that we end up consuming information and adapting certain ideas without even knowing someone is pulling the strings. Platforms have thus become an additional tool to influence people in an effective manner.

Sources
https://www.facebook.com/business/help/290009911394576

Social Engineering Defined

Please rate this

The future of digital payments: a few possibilities

30

September

2018

No ratings yet.

When it comes to payments in e-commerce, most European countries, the US, Canada and Australia strongly rely on online banking (iDeal in the Netherlands for example) Credit Card (Visa or Mastercard) payment or usage of PayPal (Adyen, 2015). The future of digital payments looks very interesting, there are many different solutions coming from all kinds of directions.

Alipay
In China, only 1% of the e-commerce payments are processed with Credit Card, whereas 48% of transactions are completed with Alipay. Unlike credit cards or PayPal, Alipay is completely integrated in Chinese lifestyles – which make it irreplaceable for a Chinese citizen. For example, you could split bills with QR codes, pay utility bills, get food delivered with WeChat, top up phone credit and buy train/metro tickets (Hendrichs, 2015). Even during Chinese New Year, people can use Alipay to send each other “red envelopes”, a traditional gift during Chinese New Year.

TechFin companies
Another new payment solutions comes from the big tech companies, with solutions such as Apple Pay and Google Pay – called TechFin solutions (Krishnakumar, 2018). These companies have enormous user bases, making it easier to penetrate to existing users of their user base. Based on their network effects, they can pose a real threat to current payment methods.

Blockchain payment methods
Also, blockchain based solutions may prove to be competition to PayPal. Solutions such as Request Network – often praised as the ‘new PayPal’ on internet fora or blogs (Levenson, 2017) – are still in its early stages. Using blockchain based solutions is also very beneficial for merchants, providing automated accounting and less need for expensive accountants or expensive accounting systems (Levenson, 2017; Yermack, 2017).

All in all, in my opinion it’s not a question if digital payments will change, but rather how it will change. With so many different new possibilities, I’m curious to see where this goes.

References:
Adyen. (2015). The Global E-Commerce Payments Guide. Retrieved from: https://www.adyen.com/blog/the-global-e-commerce-payments-guide

Hendrichs, M. (2015). Why Alipay is more than just the Chinese equivalent of PayPal. Retrieved from: https://www.techinasia.com/talk/online-payment-provider-alipay-chinese-equivalent-paypal

Krishnakumar, A. (2018). From Fintech to TechFin – Who should banks be more worried about?. Retrieved from: https://dailyfintech.com/2018/03/16/from-fintech-to-techfin-three-trends-that-banks-will-be-worried-about/

Levenson, N. (2017). Request Network is more than just PayPal 2.0 — It could revolutionize the finance world. Retrieved from: https://hackernoon.com/request-network-is-more-than-just-paypal-2-0-it-could-revolutionize-the-finance-world-87b54bb455

Yermack, D. (2017). Corporate governance and blockchains. Review of Finance 21, pp. 7 – 31.

Please rate this

The not-so glamorous side of blockchain

30

September

2018

No ratings yet.

Nowadays, blockchain seems to be on everyone’s mind. In fact, the technology behind Bitcoin has become mainstream to such an extend that most companies are exploring the implementation in their organization. By now, due to the extensive attention blockchain has received over the past years, most people will be aware of the advantages that the technology offers. Data can be stored without worrying that it can ever be altered – at least not without leaving traces – and a decentralized, trustworthy system can be created (Zheng et al. 2018). In theory, then, blockchain technology seems to be beneficial and desirable to parties that work with data, which is virtually any contemporary business.

However, as is often the case, theory and practice are far from aligned, which is evident from the 92% failure rate of blockchain projects (Maloney, 2018). First of all, the security and trust factors that characterize blockchain are founded on the use of external, independent nodes which each perform verification on a specific block of data. Although this Proof-of-Work concept does wonders for the security and trustability of data, the process itself is highly inefficient. Basically, every node will put in effort to provide the Proof-of-Work, but only one node will be rewarded. As such, the effort of every node besides that of the winner is wasted. This redundancy is further amplified as the scale of the blockchain increases and more nodes are added to the network (Back, 2018).

Another problem arises from the fact that most businesses want to develop an internal blockchain, thereby excluding any external parties to the network. As such, the blockchain will no longer be decentralized, but rather consist of several (hybrid) or only one node (centralized). Implementing such systems defeats the purpose of blockchain entirely, as the verification of data is shifted from many independent, external parties to only a few internal actors. Thus, hybrid and centralized blockchains are highly dependent on the trustworthiness of the internal ‘nodes’, thereby opposing blockchain’s core ideology to distribute consensus and responsibility among many to create trust.

Even if we put these issues to the side, one fundamental weakness remains; blockchain does not account for the quality of the input data. Sure, the data can not secretly be altered, but for data to be added to the blockchain, one simply needs majority consent from the network of nodes. Thus, if the majority of nodes believe the information is true or behave maliciously, data is added without necessarily being true and/or accurate (Bauerle, n.d.). This is particularly problematic for centralized blockchains – the version preferred by most firms – because the system’s node(s) can basically determine what information is stored on the blockchain, and what is not. This gives substantial power to the whoever controls the node and raises the question to what extent that person can be trusted, which is exactly what blockchain was designed to overcome.

In short, blockchain has great potential and can be beneficial to businesses worldwide, but the current state of the technology and its capabilities is overhyped, and its flaws often overlooked. So, would you still place your bets on blockchain to disrupt future business?

For further reading, I would highly recommend Stinchcombe’s writings (link below)!

Sources:

Back. A. (2018). The 5 major problems with bitcoin and blockchain technology. Retrieved 30-09-2018 from https://blockchainreview.io/blockchain-bitcoin-problems-limitations-issues-weaknesses/

Bauerly, N. (n.d.). What are blockchain’s issues and limitations?. Retrieved 30-09-2018 from https://www.coindesk.com/information/blockchains-issues-limitations/

Maloney, C. (2018). 92% of all blockchain projects fail. Retrieved 30-09-2018 from https://ethereumworldnews.com/92-of-all-blockchain-projects-fail/

Zheng, B., Zhu, L., Shen, M., Gao, F., Zhang, C., Li, Y., & Yang, J. (2018). Scalable and Privacy-

Preserving Data Sharing Based on Blockchain. Journal of Computer Science and Technology, Vol. 33(3)

Stinchcombe’s article:

https://medium.com/@kaistinchcombe/decentralized-and-trustless-crypto-paradise-is-actually-a-medieval-hellhole-c1ca122efdec

Please rate this

Personalized ads on TV

30

September

2018

5/5 (1)

Know that feeling seeing a really nice sweater on a website, but closing the tab because you thought it was too expensive? And then, five minutes later while scrolling through Facebook, seeing that exact sweater pop-up on the side of your feed again? We already know that when on the internet, you can find personalized ads. You can choose to opt out of these, by ticking off the right box at the cookies pop-up, that show up when opening a new website. We are familiar with these kinds of personal ads, but right now KPN is testing personalized ads on television¹.

It will be possible to show ads for certain tv-shows or documentaries, based on what you have watched before and other personal information. KPN also promises customers to see the same advertisement less often. Right now you could watch a movie on live television and see the same commercial in every break, which means about every 20 minutes. Personalized ads would reduce this, while also providing you with advertisements that might appeal more to your personal interests.

This sounds promising, however a difficulty with this is the length of the commercial break. The length of the personalized advertisements needs to be exactly the same as the amount of time planned for a specific commercial break. Otherwise, you would have different starting and ending times of the television shows, which would lead to big planning problems especially with live television. Another difficulty is gathering data to personalize the advertisements. Gathering data online is easier than gathering data from these type of customers. This might be partially solved if a television network combines online data and customer data, but that depends on a tight relationship with the customer. The growing concern around privacy is making it generally more difficult for companies to use customer data as well².

To conclude, personalized advertisements are currently being tested by KPN and have interesting possibilities for the future. However, there are some difficulties to this, which might explain why competitor Ziggo does not have any plans for personalized television advertisements (yet).

 

 

1. https://nos.nl/artikel/2252164-je-buurman-krijgt-bij-kpn-straks-andere-tv-reclames-te-zien-dan-jij.html

2. https://www2.deloitte.com/insights/us/en/industry/retail-distribution/sharing-personal-information-consumer-privacy-concerns.html

Please rate this

Artificial Intelligence in personnel selection.

30

September

2018

5/5 (9)

Nowadays, a human resource manager spends a lot of time selecting the right staff for their organization. To see if the candidates fit the job description, sufficient and trustworthy information needs to be evaluated (Dipboye 2014). Managers increasingly question the credibility of the provided information by candidates on their CVs (Weiss, Feldman 2006). Artificial intelligence can come into play assisting human resource managers and there is even a possibility of making them obsolete in the future.

Unilever, a transnational consumer goods company, already started experimenting with staff selection assisted by artificial intelligence making resumes superfluous. The company starts by scanning LinkedIn profile data using an algorithm which drops half of the candidates (Thibodaux 2017). Subsequently, several games using artificial intelligence to assess and match candidates to the company have to be played. Finally, less than the top third submits a video interview focusing on business challenges (Gee 2017). All these steps combined will accelerate the human resource pre-selection phase without the intervention of humans. Furthermore, according to Unilever the hiring process has become more accurate as 80 percent of the applicants in the final round are offered a job (Gee 2017).

Although these developments in the selection procedure may seem revolutionary in a sense that the process becomes more accurate. Of course, the other side of the coin should also be considered. The use of artificial intelligence in the process also raises some questions concerning the legitimacy of the selection. Artificial intelligence is not entirely unbiased as it is basing its decisions on data provided by humans. Besides, is it ethical to exclude people based on a decision made by artificial intelligence? As more people are selected through this process the companies behind it should measure the effects. Think of cases where too many people with the same background are selected or it influences company performance negatively. As artificial intelligence has already made its way to the pre-selection phase of personnel in a big company like Unilever, people should start thinking of the consequences of using it.

Dipboye, R.L., 2014. The role of communication in intuitive and analytical employee selection, Meeting the challenge of human resource management: A communication perspective 2014, Routledge New York, NY, pp. 40-51.
GEE, K., 2017. In Unilever’s Radical Hiring Experiment, Resumes Are Out, Algorithms Are In. Dow Jones Institutional News.
THIBODAUX And WANDA, June 28th, 2017-last update, Unilever Is Ditching Resumes in Favor of Algorithm-Based Sorting. Available: https://www-inc-com.eur.idm.oclc.org/wanda-thibodeaux/unilever-is-ditching-resumes-in-favor-of-algorithm-based-sortingunilever-is-di.html?cid=search [27-09-, 2018].
Weiss, B. and Feldman, R.S., 2006. Looking good and lying to do it: Deception as an impression management strategy in job interviews. Journal of Applied Social Psychology, 36(4), pp. 1070-1086.

Please rate this

Adverse Effects of AI

30

September

2018

5/5 (4)

Over the past decade, the field of artificial intelligence (AI) has seen fascinating developments. There are now over twenty domains in which AI programs are performing at least as well or even better than humans. While AI can be used for our benefit in many areas, it also can be misused for malicious ends.

There are many risks to the use of AI, as the risks associated with privacy and data security are real. As AI enhances the expected value of data, firms are encouraged to collect, store and accumulate data, regardless of whether they will use AI themselves (Zhe Jin, 2018). The ever-growing big data storehouses become a prime target to hackers and scammers. A concrete example of harm that could arise from a data breach is identity theft. Scammers were engaging in this area long before big data and AI existed.

Recent trends suggest that criminals are getting more sophisticated and are ready to exploit data technology. For instance, robocalls – the practice of using a computerized auto-dialer to deliver a prerecorded message to many telephones at once – has become prevalent because of relatively standard advances in information technology (Burton 2018).

However, as a result of AI, criminals are using improved methods of pattern recognition and delivery, increasing their efficacy of these calls, e.g. The receiver of the call will see a local number that looks familiar to him, this could even be a number from his or hers personal contacts. The call is then tricking the receiver into listening to unwanted telemarketing.

Another matter around AI and other predictive technology is that they are not fully accurate in their intended use. While it may not introduce much wasteful effort is apps like Netflix cannot precisely predict the next movie people want to watch, it could be much more consequential if the U.S. National Security Agency (NSA) flags certain innocent people as a possible future terrorist based on shortcoming of an AI algorithm (Zhe Jin 2018).

To summarize, there is a real risk in privacy and data security. The magnitude of the risk, and its potential harm to consumers, will likely depend on AI and other data technologies. What are your thoughts on these risks?

References:

Burton, J. (2018). Hacking your holiday: how cyber criminals are increasingly targeting the tourism market. [online] The Conversation. Available at: http://theconversation.com/hacking-your-holiday-how-cyber-criminals-are-increasingly-targeting-the-tourism-market-98967 [Accessed 30 Sep. 2018].

Zhe Jin, G. (2018). ‘Artificial Intelligence And Consumer Privacy’, in NBER Working Paper 24253, pp.4 – 8. Cambridge: National Bureau of Economic Research.

Please rate this

Taking Care of Virtual Patients

30

September

2018

No ratings yet.

1*l9qYtN7RcBOk9ontE6Nm4A

In engineering, the concept of ‘Digital Twins’ has gained attention during the last decade. A Digital Twin is a virtual representation of a physical object, which is continuously fed with data from embedded sensors and software. Hereby, the Digital Twin tightly connects the physical system with its computer model. Digital Twins are, for examples, widely used to continuously monitor and forecast the health of jet engines. This allows airlines to identify how and where potential problems could occur, whereby predictive maintenance is deployed to keep the system healthy.

In this blog, I will explain what the possibilities and corresponding benefits and risks are for this technology in the healthcare industry. The enhancement of computational power and molecular readout technologies has increased the potential of ‘virtual patients’ to continuously track health and lifestyle parameters. As of now, the digital models used in healthcare are quite partial (such as twin models of the heart) and basic. Yet, already signs of the effectiveness of these models can be observed as well as the benefits it could bring in the future.

First, the data-rich Digital Twins would allow for the creation of a more detailed picture of the patients which results in faster and more accurate identification of actual or potential disease states. Hereby, a shift to more preventive solutions could result in significant health improvement and hence reductions of health care costs. Second, the multidimensional properties of the digital twins could allow practitioners to more accurately compare a patient’s health with the health stats of similar patients. Since clustering can be based on more elements than, for example, age and gender, deviations from the ‘normal’ can be identified faster and more accurately.

Yet three main societal concerns are also worth noting. First of all, Digital Twins could raise inequality since developing a digital version of yourself could be very costly. Hence, the benefits of improved health and possible life extension could potentially only be accesses by wealthy people. Second, the Digital Twin could lead to self-fulfilling prophecy mechanisms where knowing that you could potentially become sick in the future will make you indeed feel sick and weak. Thirdly, it is of great importance to ensure data protection. Data leaks could quickly offset the potential benefits of Digital Twins, as for example, insurance companies could use the data to modify the insurance policies for individuals in their favor

The future will tell whether we will be able to effectively govern this emerging technology in the healthcare industry; thereby significant health and cost benefits can be obtained by actively managing the associated concerns.

 

 

 

Sources:

 

Bruynseels, K., Santoni de Sio, F., & van den Hoven, J. (2018). Digital twins in health care: Ethical implications of an emerging engineering paradigm. Frontiers in genetics, 9, 31.

 

Mussomeli, A. (2018). Expecting Digital Twins. Deloitte Insights. Retrieved from: https://www2.deloitte.com/insights/us/en/focus/signals-for-strategists/understanding-digital-twin-technology.html

 

Van Houten, H. (2018). The Rise of the Digital Twin: How Healthcare Can Benefit. Philips Research. Retrieved from:https://www.philips.com/a-w/research/blog/the-rise-of-the-digital-twin-how-healthcare-can-benefit.html

 

 

Please rate this

Sharing Data With Apps: Does It Matter?

30

September

2018

5/5 (5)

You’ve just downloaded a new app and are starting it up for the first time. After the loading screen a bunch of questions pop up: can X-app send you messages? / share your location with X-app / give X-app access to your contacts-photos. It never ends. Everyone is familiar with this scenario, but what does it matter?

Although it doesn’t seem dangerous to share these details, it might put you in more harm than you think. Sharing your location can expose yourself, or others, without you wanting to or knowing it. When using car sharing apps like Uber or Lyft, or tracking your exercises with Apple watch or Strava, you’re trusting these companies with information about yourself. Unfortunately, these companies don’t always properly take care of the data. 4iQ monitors the surface, social and deep and dark web for identity records exposed in data breaches and accidental leaks1. They found ride sharing companies in Mexico and India that accidentally exposed sensitive information to the web. User’s ride requests include the time, exact pick-up location, number, addresses etc. This information and more was all available with a bit of digging.

When this information is combined with other location-based services, or even Twitter and Instagram, it can impose a real threat. The site PleaseRobMe.com is a prime example of this. It combines a stream of updates from various location-based networks and shows when users have checked in somewhere with for example Instagram… and thus aren’t at home. Knowing someone isn’t home gives the perfect chance for burgers to go and rob them (Siegler, 2018).

These cases already show the importance of keeping your privacy in check, but Strava stepped up the game in giving away information that probably shouldn’t be given away (Blue, 2018). Classified information even. In 2017, Strava (an app that tracks your exercises, where you’ve been, how fast etc.) published their global heat map. This heat map was built up from 1 billion sportive activities, 3 trillion longitude and latitude points, and 10 terabytes of data. It shows the most used trials to run, or best roads to ride your bike. It does, however, also show the location and patrolling routes of military bases, like this one in Kandahar, Afghanistan (Triebert et al., 2018).

So, willingly sharing data like your location could make yourself a target of criminals with malicious purposes. Or when combined with millions of other locations, it can even lead to the military reviewing their guidelines for wireless devices (Sly et al., 2018). These things definitely make me think about the privacy settings on my phone and whether companies are properly taking care of my data.

References:

  1. https://4iq.com/
  2. Siegler, M. (2018). Please Rob Me Makes Foursquare Super Useful For Burglars. [online] TechCrunch. Available at: https://techcrunch.com/2010/02/17/please-rob-me-makes-foursquare-super-useful-for-burglars/ [Accessed 30 Sep. 2018].
  3. Blue, V. (2018). Strava’s fitness heatmaps are a ‘potential catastrophe’. [online] Engadget. Available at: https://www.engadget.com/2018/02/02/strava-s-fitness-heatmaps-are-a-potential-catastrophe/?guccounter=1 [Accessed 30 Sep. 2018].
  4. Triebert, C., Koetll, C. and Tiefenthäler, A. (2018). How Strava’s Heat Map Uncovers Military Bases. [online] NYTimes.com – Video. Available at: https://www.nytimes.com/video/world/middleeast/100000005705502/big-data-big-problems-how-stravas-heat-map-uncovers-military-bases.html [Accessed 30 Sep. 2018].
  5. Sly, L., Lamothe, D. and Timberg, C. (2018). U.S. military reviewing its rules after fitness trackers exposed sensitive data. [online] Washington Post. Available at: https://www.washingtonpost.com/world/the-us-military-reviews-its-rules-as-new-details-of-us-soldiers-and-bases-emerge/2018/01/29/6310d518-050f-11e8-aa61-f3391373867e_story.html?noredirect=on&utm_term=.798cbfa5ea54 [Accessed 30 Sep. 2018].

Please rate this

“Say hello to our building.” – When cognition meets architecture

30

September

2018

No ratings yet.

“Good Morning, Nicole. I hope you had a pleasant weekend. I will now direct you to today’s workspace – please follow me. Temperature and lightning have been adjusted to your preference. I also noticed your vehicles low battery level – I will make sure to connect it to the charging station. Your flight to Frankfurt departs at 3 PM today – a cab will pick you up at 1.30 PM.”

One might think I have a very attentive and organized assistant, right? However, it is 2020 and not a human greeting me so friendly – the building is talking to me. Say hello to the future of work.

At the beginning of the century, automated buildings began to transform into smart buildings. They learned to measure and understand energy consumption and asset utilization levels, but those processes were generally constraint to the analysis of primary data points – no two-way integration, no learning, no reasoning.

Thanks to Moore’s Law withstanding and progress in IoT and sensor technology exponentially growing, a new game changer is entering the field – welcome to the era of cognitive buildings. The complex practice of merging architecture with highly advanced technology creates the next generation of sustainable building systems. This allows us to not only eliminate prior limitations with regard to analytics, but to entirely redefine the meaning and processes of work and organizational culture. Leveraging sensor technology in combination with deep learning mechanisms allows the building to scale the deployment of thousand data points, finally anticipating user behavior, driving down cost, enabling innovative collaborative services and – last but not least – reduce the carbon footprint (EnOcean Blog, 2017).

Now, consider the bigger picture: industry specific application of cognitive buildings, such as in hospitals or airports, will finally create entire networks of interconnected cognitive buildings, transforming the way we live significantly (Harrison, R. 2016). However, this clearly creates risk with regard to privacy and data security and the general uncertainty that comes with advancement – but then, when did the fear of uncertainty ever stop human kind from progress?

Sources:

EnOcean Blog. (2017). Cognitive Buildings. EnOcean Blog. Retrieved from:
www.blog.enocean.com/cognitive-buildings/

Harrison, R. (2016). Cognitive Buildings: Smartening Up for the Future. GenslerOnWork. Retrieved from: www.gensleron.com/work/2016/7/8/cognitive-buildings-smartening-up-for-the-future.html

Bloomberg (2015). World’s Greenest Office Building Is Dutch: The Edge. YouTube, Retrieved from:
www.youtube.com/watch?v=JSzko-K7dzo

Please rate this

AI as your new best friend

30

September

2018

5/5 (1)

Think about how much time you spend on your social media each day. Social media was expected to facilitate connectivity between people and be a solution for isolation and loneliness. However, the opposite has happened (Bird 2018). Nowadays, people are spending more time on their phone than actually socializing. Moreover, social media is mirroring a perfect world, which can lead to people comparing their lives and feeling dissatisfied.

Experts are now looking at possibilities of using AI in order to counter this negative effect. Additionally, AI could help with getting more people treated. The amount of people suffering from a mental illness is growing, yet only a small group is getting the treatment they need. Waiting lists for therapy sessions are long and other people do not seek help due to the costs. However, currently AI is mostly a supporting system for mental healthcare. AI can assist current therapists, making it possible for therapists to treat more people (Stix 2018).

An example of AI used in mental healthcare is ‘Tess’, which is a mental health chatbot. Tess is an instant messaging application through WhatsApp, Facebook Messenger, SMS and web browsers. Tess provides personalized mental health care services, such as cognitive behavioural therapy, coaching and psychotherapy. She evaluates how people are feeling with an emotion algorithm and recognizes when there is a downward trend. The emotion algorithm is combined with a natural language processing algorithm, enabling Tess to understand what a person is talking about (Berman 2016). Moreover, Tess does not use pre-selected responses, she is designed to react to shifting information. She remembers what people say and is able to mention it again at a later time. When she thinks the mental illness becomes more severe, she will connect you to a therapist (Gionet 2018)

Tess is available at all times and users do not have to pay an hourly fee, which could lead to many more people seeking help. However, it still has its practical limits; personal contact and the connection between patients and therapist is an important aspect of the success of a treatment.

The question remains what effect these chatbots will have in the future, when AI might not be used as a ‘’partner’’ of a therapist, but as the therapist itself.

 

References:

Berman, A. (2015). Bridging the mental healthcare gap with Artificial Intelligence. [online] Singularityhub. Available at: https://singularityhub.com/2016/10/10/bridging-the-mental-healthcare-gap-with-artificial-intelligence/ [Accessed 30 Sep. 2018].

Bird, R. (2018). AI soon to be your best friend and mental health therapist?. [online] Hewlett Packard Enterprise. Available at: https://www.hpe.com/us/en/insights/articles/ai-soon-to-be-your-bff-and-mental-health-therapist-1802.html [Accessed 30 Sep. 2018].

Gionet, K. (2018). Meet Tess: The mental health chatbot that thinks like a therapist. [online] The Guardian. Available at: https://www.theguardian.com/society/2018/apr/25/meet-tess-the-mental-health-chatbot-that-thinks-like-a-therapist [Accessed 30 Sep. 2018].

Stix, C. (2018). 3 ways AI could help our mental health. [online] weforum. Available at: https://www.weforum.org/agenda/2018/03/3-ways-ai-could-could-be-used-in-mental-health/ [Accessed 30 Sep. 2018].

Please rate this