Last updated: 06 September 2019
As the digital transformation of businesses and services continues with full force, artificial intelligence (AI) has become somewhat of a buzzword in the technology sector. While it’s true that we haven’t quite reached the level of technology sophistication often shown off in Hollywood blockbusters, there already are a variety of use cases where machine learning algorithms are being deployed to improve different aspects of our daily lives. Below, we look at four industries that are reaping the rewards of using AI and what this might mean for the future.
Healthcare is one of the most promising areas likely to be transformed significantly by AI and machine learning. This is because this technology can quickly go through large amounts of data and find patterns that humans might miss. There have been various use cases where AI has been able to diagnose patients more accurately and speedily than medical experts, allowing doctors to treat their patients before their condition worsens. For example, AI has been used in a trial to treat atrial fibrillation, a condition of the heart that increases the risk of stroke. Computing modelling was found to be able to detect subtle signs, such as scarring of the heart, which are unable to be spotted by the human eye from other test results. The computer system analyzed tests carried out on nearly 181,000 patients between 1993 and 2017 who all had had normal test results at first. The modelling correctly identified the subsequent diagnosis from the normal test results in 83% of cases.
In a similar case, AI deployed by DeepMind Health was able to warn doctors that their patients were at risk of developing severe kidney damage. The AI analyzed very subtle changes in patients’ conditions and was able to correctly predict if the disease developed in nine out of 10 patients. It is also worth noting that the AI could foresee the condition emerging up to two days ahead of doctors. It is clear that AI has a huge potential to save lives, although it’s important to note that this research is still in its early stages.
If you’re interested in more examples around how AI and cutting-edge data analysis is being implemented in a range of ingenious ways in healthcare, from developing new medications to helping prevent and treat diseases, read our article on the topic.
Travel and Tourism
In previous blogs we have discussed the advantages of using AI to facilitate the airport experience for passengers. Our Fly to Gate technology, for instance, uses facial recognition to authenticate a passenger’s journey all the way through the terminal to the boarding gate to create a fully self-serviced journey. However, other areas of the airport experience have also started to leverage AI. In August 2019 for example, University of Arizona spinoff Discern released its virtual border guard that doubles as a lie detector by using AI to spot anomalies in responses. The Avatar asks travelers a series of questions and analyzes information including facial expressions and tone of voice to search for deception signals.
The intent of the system is to filter out people with dangerous intentions who are trying to cross borders. And with a lie detection success rate between 80%-85%, the system was much more accurate than humans trying to distinguish the truthfulness of the same statements, who averaged 54%. Using AI to help human guards at airports enhances passenger safety, reduces security risks and decreases their current workload to focus on truly suspicious suspects, providing another layer of defense against threats.
You may have heard of the viral FaceApp challenge – a mobile app that uses AI algorithms to age users by 30 years. And while there is scope for AI in cases like this to be used on social media in a lighthearted manner, there are much more serious problems on social media that AI can help tackle. One example is the trial by Hatelab and Samurai Labs who are using the technology to help monitor aggressive anti-Polish hate crimes on social networks. Specifically, AI categorizes the different levels of abuse, helping authorities recognize when there is a legitimate and real threat to someone’s life.
Additionally, Instagram and Facebook have put in place AI programs to tackle various problems on their networks. Instagram has begun using DeepText AI in order to try and curb cyberbullying by identifying negative comments before they are published. This program was originally used in 2016 to seek out spam and has since been trained to also block racist comments and inappropriate photos.
Facebook has taken a different approach and used AI software to scour linguistic red flags that could be a sign of depression. The program was proven to be very effective, identifying the warning signs of depression up to three months before medical health professionals diagnosed the condition. Again, this idea is in its early stage of development but with the ever–increasing pressure on social networks to tackle mental health problems associated with them, AI could be a pivotal part of the solution.
The hospitality sector has also been in the spotlight for using AI, with a bar in London becoming the first ‘AI bar’ in the World. The venue uses a facial recognition system that puts customers in an intelligently virtual queue, helping the bartenders ensure that they serve drinks on a first come first serve basis. This has increased convenience and efficiency both for staff and for customers. In addition, the system has also been used to help speed up ID checks at the bar. For example, if a customer looks underage the system will prompt them to have their ID ready or let the bar staff know the ID document has already been checked at the door.
In the future, the company hopes to add several other features, including the ability to set up a bar tab based on your face and to order drinks while still in the queue. It will be interesting to see if this idea inspires similar establishments to begin using AI in this way.
As machine learning technology develops further, it is exciting to look at the industries where it is making leaps forward. Not only is AI starting to make our lives more efficient, it is also being used to provide convenience and even save our lives.
However, more work still must be done to ensure that humans creating the AI algorithms have a wide variety of data and diversity so that AI can perform accurately. AI systems are only as good as the data we put into them. Bad data can contain implicit racial, gender, or ideological biases. There have been several cases in which AI has made wrong judgements based on gender, race or sexuality and biased AI systems are likely to become an increasingly widespread problem as AI moves out of the data science labs and into the real world. But if we can measure the likelihood of bias in the AI systems we create and attempt to remove our own biases in the data we input, then AI could be used in even more cases. Harnessing this intelligence to create change is an exciting prospect.