AI is listening to you

Recent advances in voice recognition bring AI technologies to the home

By Natalie Mouyal

Artificial Intelligence (AI) was one of the big buzzwords at CES 2018. From home appliances to robots and self-driving cars, AI is able to help us with our everyday activities. While an interest in intelligent machines can be traced back to Greek mythology, recent advances in computing that enable us to collect large quantities of data and then process it using algorithms, have hastened the development of AI technologies.

AI robot touching screen Recent advances in computing have hastened the development of AI technologies

AI technologies rely on IEC Standards

The term artificial intelligence is generally understood to refer to a machine that can replicate cognitive functions such as learning and problem-solving. It is a broad concept that encapsulates ideas ranging from Frankenstein-like robots to voice assistants for smart phones and other devices.

The Joint Technical Committee of IEC and ISO on information technology (ISO/IEC JTC 1) and several of its subcommittees (SCs) prepare International Standards which contribute towards artificial intelligence. Given the rapid developments in AI across many industries, a new SC on artificial intelligence, ISO/IEC JTC 1/SC 42, was set up in 2017 with the mandate of providing standardization in the area of AI as well as guidance to other committees developing AI applications.

AI depends on the gathering, analysis and sharing of great volumes of data which are exchanged between applications as well as with external service providers. This makes it equally possible for an assistant-enabled device to turn on the oven or for a manufacturer to improve the safety features on its cars. ISO/IEC JTC 1/SC 41 develops International Standards for the internet of things (IoT), making connectivity possible, while ISO/IEC JTC 1/SC 38 addresses the standardization of cloud computing for the storage and retrieval of data.

In addition, AI technologies rely upon IEC Standards for hardware components such as touchscreens (IEC TC 110) and audio, video and multimedia systems and equipment (IEC TC 100).

The latest AI technologies at CES

At CES this year, assistant-enabled devices using voice recognition were ubiquitous. Improved voice recognition functionality and integration into home appliances means you can now ask your refrigerator to pull up recipes and order groceries online or your television set to change channels.

Increasingly, car manufacturers are incorporating voice recognition into their models. Voice commands can be given to modify the interior temperature or play a certain song. In one model, the car adapts to passenger preferences by, for example, suggesting directions to a favourite restaurant at dinner time. These applications rely upon the work of ISO/IEC JTC 1/SC 37, which develops International Standards for generic biometric technologies including voice recognition. In addition, Technical Area (TA) 16 of TC 100 (IEC TC 100/TA 16) addresses the issue of voice recognition within the context of active assisted living (AAL).

CES was also showcasing the latest AI technology being used to help self-driving cars understand their environment and react accordingly. Smart sensors are an essential component, allowing for the collection of the data required for vehicles to adjust their driving based on factors such as their location, the driving conditions or vehicles nearby. Other smart sensor features, such as lane-keeping, blind-spot monitoring and automatic braking systems, are already in use.

These technologies are aided by the work of IEC TC 47, which develops International Standards for the design and use of semiconductor devices, including sensors. IEC SC 47F specifically addresses microelectromechanical systems (MEMS) which are used for automotive applications, while IEC SC 47E provides International Standards for sensors used for imaging, motion and distance detection. 

Other technologies used by autonomous cars are also addressed through the work of the IEC. TCs standardize on technologies such as dashboard touchscreens (IEC TC 110), cloud computing (ISO/IEC JTC 1/SC 38) for processing, analyzing and storing data, and near field communications (ISO/IEC JTC 1/SC 6). Work within IEC TC 100/TA 17 covers the areas of multimedia systems and equipment for cars and recently resulted in the release of Technical Standard IEC TS 63033, which enables drivers to spot obstacles using an intelligent wrap-around view monitor.  

The market for AI technologies used in cars is on course for rapid growth. According to analysts at the research company IHS, unit shipments of AI systems are expected to rise from seven million units in 2015 to 122 million by 2025.

Keeping AI safe and secure

While the introduction of new AI technologies has generated much excitement, it also raises concerns over security and the protection of data.

The ubiquity of connected devices that are able to communicate with each other increases the number of gateways that can potentially be used to breach a system, whether at home or in a factory or car. Cyber attacks can have disastrous consequences, causing production in a factory to shut down or a home to be robbed.   

Data protection is becoming increasingly important as connected devices collect vast amounts of information about their users on a daily basis. Within a home, these devices can store details such as favourite songs and television shows, but also the times of day when the home is empty. This raises considerable concerns regarding privacy that will need to be resolved.

International Standards are essential tools in the battle to provide information security and protect data against a cyber attack. ISO/IEC JTC 1/SC 27, IT security techniques, has developed the ISO/IEC 27000 family of International Standards for information security management systems (ISMS) to enable organizations to keep their data assets secure. In addition, the IEC has developed industry-specific Standards to help secure critical data. For example, IEC TC 62 provides Standards concerning the electrical equipment, electrical systems and software used in healthcare. 

More industries are using AI

According to YouTube Chief Product Officer, Neal Mohan, AI accounts for 70% of the content viewed on the platform. Machine learning algorithms enable huge amounts of data to be processed and interpreted to provide patterns on which to base a prediction. As a result, YouTube can make recommendations based on our previous selections. The same approach is employed by shopping sites as well as video and music streaming platforms like Netflix and Spotify.  

Deep learning is the most recent AI technique to find real world applications. It mirrors the neural networks of the human brain to create transistor connections that can be strengthened or weakened depending on whether the data is interpreted correctly. As new data is received, the machine is trained automatically to perfect its predictions. As a result, machines are now able to recognize and respond to images and voices but can also beat human competitors in games such as Go and chess.

Algorithms are used extensively in the finance industry where it is estimated that high frequency AI trades account for more than half of equity share transactions on the US market. AI technology is also used to help detect fraud and communicate with customers via chatbots.

In the manufacturing sector, AI technology provides data intelligence and automation. Processes are automated to increase efficiency while data analytics and predictive algorithms enhance operations and strategy. Even cyber security benefits from AI technology in helping detect and defend against attacks.

Applications in the medical field using AI technology can detect anomalies in images and undertake semi-automatic tasks during minimally-invasive surgery. In the future, AI technology will mine through data to determine a patient’s risk for certain diseases or provide an initial diagnosis. AI will also enable the development of individualized therapy based on the genetic profile of the patient.  

Voice-recognition assistants, relatively commonplace in smartphones, are increasingly found in homes as a hub through which to connect and control household appliances, order products online and stream music. Other examples include thermostats that regulate the temperature depending on whether anyone is home and security cameras that use facial recognition to decide whether to unlock the front door. These technologies rely on IEC Standards for sensors (IEC TC 47), cloud computing (ISO/IEC JTC 1/SC 38) and biometrics (ISO/IEC JTC 1/SC 37).

In the near future, cities will be using AI technology to improve safety, traffic management and infrastructure maintenance.

Looking ahead

It is not yet clear how profound an impact AI technology will have on our lives. While one Silicon Valley CEO has compared its importance to the discovery of electricity and fire, it is not clear if this is truth or hyperbole.

While we are still many years away from machines that are as smart as – or smarter than – the human mind, technical advances are progressing rapidly. The IEC continues to follow the latest advances closely and develop the relevant International Standards.

Gallery
AI robot touching screen Recent advances in computing have hastened the development of AI technologies
In-car voice recognition system AI brings voice recognition inside cars
Landroid folds your laundry And helps fold your laundry (Photo: Landroid)