AI Use Cases 

E-records in health systems are a growing set of critically important documents. Machien learning enables them to be created more efficiently and then updated and (where necessary) consolidated more effectively and cheaply.

Across large text datasets - for example, historical archives - there will be trends or insights that can be analysed but which it would take a human observer too much time to process. Although still at a relatively rudimentary stage this will become an increasingly powerful tool for creating new insights and levels of knowledge.

Analyse data feeds about incoming cyber activity as well as behavioural data inside an organisation - married with up to date information about the broader cyber security situation - to predict potential security risks.

A key part of hacking or similar scams such as phishing is building a larger database of potential targets and a better understanding of their profiles. One AI technique to support this is to use facial recognition across platforms to identify multiple accounts. Depending on regulatory data rights similar techniques can be used for more conventional marketing.

Utilise learning systems to effectivel and swiftly respond to security threats, many of which may have been delivered with machine learning support. This is a game where every innovation in defence triggers the next innovation in attack - and vice versa.

Provide personalised learning programmes - with regular testing / feedback loops used to assess and deliver against topics that individual students find more challenging and a pace, and potentially style, of interaction tailored to the individual.

Support 3D CAD (Computer Aided Design) software - e.g. Autodesk Dreamcatcher - to build options in to the design process. Environmental, traffic data used to help analyse potential outcomes. Architects / design staff can make educated decisions with visualised data-rich options.

Breaking security systems is a growing field of study by a variety of actors - often of the malignant sort. Whether predicting PIN numbers from acoustic signals or spoofing iris images this is a race between attack and defence.

Sensors can be expensive, hard to maintain or simply unavailable for what can be important data. Using data from other sources can enable the predictive modelling of other data sets. Note that this can create a series of new risks, especially if historic patterns break down or feedback effects occur.

Ensure high quality record keeping by transcribing key medical conversations - lowering risk of potential errors due to (famously) poor medic handwriting. Provides detailed, comprehensive information capture.

Use deep neural networks to remove clutter from images - this can vary from blurring to watermarks. This will often be used when poor quality image data has been captured and mass data cleaning will help improve training data for other AI applications.

Detect potentially dangerous electrical current surges. This potentially enables systems and devices to automatically react in an appropriate fashion to minimise risk to users and devices.

Taking a 2D image this technique allows for the creation of a 3D model of the objects captured in the photos. This obviouasly works better on images of objects with known characteristics - like humans for example.

In a world where information is at anyone's fingertips plagiarism is a constant challenge, especially in academic circles. NLP can be used to match content and identify suspect material.

Monitor application performance for cloud compute resource allocation - scaling up or down and switching on or off capacity in response to web and mobile application usage numbers.

Leveraging 3rd party data to better understand existing customers and target potential customers is critical. However, the basics of ensuring correct data matching among different sets is essential to deliver this and can be very labour intensive.

Deploy decoy-assets in a network as bait for attackers, to identify, track, and disrupt security threats. The aim is to lure attacks in to information traps to allow for an augmented security response.

"Codifying business expertise so that is more widely available and understandable. For example investment decision-making approach or economic modelling skills

Automate project management to monitor progress and identify risks

Detect anomalies in software stacks (e.g. security or downtime risks). Fro example, banks tend to have complicated legacy systems based both on multiple generations of development and software codes but also driven by historic merger activity in the industry.

Searching for data from 3rd party sources on websites to capture, and potentially quantify, knowledge

Update and improve corporate Intranet content through auto-edited and standardised content tags - for example ensuring that engineers have up to date sample code and best practice guidelines collated on a standardised basis. User searching is used to improve the data tagging.

Legacy systems can be a significant challenge in the process of increasing efficiency and effectiveness of organisations. There may well be a multitude of such systems with limited inter-operability. Deploying machine learning (ML) and robotic process automation (RPA) can ensure that these systems can become part of a newly effective process - reducing the new for system replacement and streamlining processes. Platforms are being developed to do precisely this.

Using wifi signals to identify which individuals are moving - even through physical barriers such as walls. Potential real world uses will include security or facilities tasks - and in the medium term various healthcare or interactive gaming scenarios.

Build on feedback from residents to create optimal mood-related visual stimuli in residential or office settings. This may include support for better interior design (using associated technologies including augmented reality) but also may lead to annoying whale music.

Optimise labour staffing and resource allocation to reduce bottlenecks in medical institutions - in many hospital systems significant time for example is spent researching into, and managing, bed availability.

In a world of growing data availability and machine learning medical practitioner performance indicators can be better measured. These can include overall outcomes as well as individual actions (e.g. prescription practices) and the data can be more easily standardised.

Physician attrition is a consistent issue in most medical systems given the innate pressures of the role. Downsides for the individual (emotional, physical, mental), their patients and their employers all need to be avoided.

Using robotic sensors to capture surgeons' performances during robot-assisted surgery and analysing resulting performance data. This has a knock-on impact of potentially reducing medical malpractice liability exposure.

Monitoring voice calls by traders to test for signs of wrong-doing Intelligent Voice have the potential to deliver this service.