Innovation in artificial intelligence (AI) is turning what was once science fiction into reality. With the exponential pace at which technology develops, barriers are being broken in AI faster than many have thought possible. On 30 November, AER’s event Artificial Intelligence: Are Regions Up to The Challenge? will bring together field experts, academics, and stakeholders giving attendees the opportunity to learn about the latest trends and innovations in AI.
A Tool of the Present and Future
Discussions around AI often involve looking far into the future where the world is inhabited by autonomous robots, sometimes minimising the large role that AI currently plays in our lives. While like Google’s Alpha Go beating Master Go player Ke Jie or Tesla’s self driving cars may attract the headlines, innovation in AI is ongoing and has produced many other inventions which are changing the nature of longstanding industries. Being attentive to these more subtle present day changes and continuous innovation in AI allows people to predict the direction each industry may take in the near future. Understanding the trends in each industry paints a more accurate picture of the roles AI will come to play in society, giving policy makers some of the knowledge needed to help integrate AI into society.
While the abilities of smartphone assistants like Apple’s Siri, Facebook Messenger’s bots, and Google’s Assistant have long been chronicled, closely related home assistants are proving increasingly capable themselves. Amazon’s Echo and Google Home can already switch lights on and off, set timers while people cook, and play songs or answer questions on demand. Other tasks that each can do range from telling users about nearby restaurants, checking people’s calendars for them, and playing trivia games. For now little separates phone assistants from smart homes but tech giants future plans may change this. With landlines becoming obsolete, Amazon and Google are positioning their devices to take its place. Both companies are investing in voice-calling features for their devices, with their goal of making digital phone calls like Skype completely hands free. Additionally, to improve the functionality of their devices the companies are integrating their smart homes with more and more other smart devices. The ultimate vision of smart homes is to make everything from adjusting the thermostat, turning on and off lights, to locking doors all integrated under one system. Such a system would give people the power to not only do these things with a simple command but all the way from another continent as well. One of the most promising uses of smart technology is helping elderly citizens with active and healthy ageing, a topic with AER has been active on in the past. Smart homes are already being employed in ambient assisted living which utilises smart sensors and other wearable devices to capture data about patients’ daily activities and present state of health, making it easier for elderly citizens to live in their own homes as the age.
Smartphone users are already familiar with machine learning using personal data to automatically add events to their calendar and planning the best route home based on current traffic conditions. Now, one of the next frontiers for machine learning will be integrating itself into the daily lives of consumers to provide personalized online shopping. One app, called Pinterest Lens, lets users take a photo of a desired object and finds the item online. Another image detection algorithm can identify an object and searches for similar ones online, making it easier for consumers to buy a desired product. These apps offer a glimpse into the future of predictive shopping where programs and smart devices will be able to mine massive amounts of data ranging from “consumers’ purchase histories, product preferences, and schedules; competitors’ pricing and inventory; and current and forecasted product demand” to give consumers a highly personalized shopping experience.
When people are running low on Tide laundry detergent they can already press Amazon’s wireless Dash Button for it to order more through Amazon Prime. As predictive shopping evolves it will be able to learn users’ behavioural and environmental habits to produce more personalized options. The more data that retail store’s programs collect the better they will become at determining customers’ specific needs. Certain stores are currently using smartphones to pick up on consumers’ behaviour and provide them with context-specific recommendations. Soon, it is very possible that an appliance such as a Samsung Smart Fridge will be able to calculate when users are running low on certain foods and notify them in the morning before they leave for work. Furthermore, in the not too distant future it is likely that retailers will be able to deliver personalized content according to what mood someone is in, the activity they just finished, how much time they have to shop, or the time of day it is. Going forward, beyond developing the right algorithms challenges for retailers will include the protection of consumer data and being subtle enough to avoid giving users the sense that their personal lives are being intruded upon.
One of the most rapidly progressing areas of AI is in multimedia. Right now, using AI people can already create 3D face models from a 2D image, generate sound effects for a silent video, and insert smiles onto any celebrity using Twitter bot Smile Vector. These are just some of the AI assisted multimedia tools people now have at their disposal and new technologies are developing at a breakneck pace. In just one year a program went from producing pictures such as this to this. While these advances in multimedia could be a boon to the creative arts industry, they will almost certainly have negative side effects as well. With society already having problems combating the proliferation of misinformation online giving nearly anyone the ability to create believable images and videos could exacerbate the issue.
AI and Agriculture
Since humans shifted from a hunter gather lifestyle to a more sedentary agrarian lifestyle, farmers’ crops have been plagued by crop diseases. In the past, pathogens have destroyed everything from the world’s most popular coffee beans to people’s preferred variety of banana. Today, nearly 40 percent of crops are annually lost to diseases. Fortunately, the latest developments in AI may decrease that figure dramatically by improving the accuracy in diagnosis of crop disease. At the forefront of the industry is a team researchers from Pennsylvania State University in the U.S. and the Ecole Polytechique Federale in Switzerland who have created a program which will detect crop diseases before they spread. By giving the program massive datasets-over 50’000 photos-the researchers taught it to detect 26 diseases in 14 plant species with 99.35% accuracy. Unlike many other AI programs where a drawback is the program’s price point, the researchers’ program will enable anyone with a smartphone to take a picture of a crop and receive a diagnosis in seconds. As it moves past its trial phase, the next step for researchers will be improving the program’s ability to detect diseases in a wide range of settings.
Advances in the agricultural industry due to AI are not just limited to the diagnosis of crop disease, but can now play a role in the treatment of them as well. Another machine learning program, this time an automated tractor, can identify weeds in lettuce crops within 0.2 seconds and spray them with great precision. Project worker Ben Chostner claims that farmers can reduce their pesticide usage by 90 percent using his company’s robots. If rolled out successfully, the program could transform how farmers deal with threats to their crops. The current approach to using pesticides is akin to “carpet bombing fields”, causing chemicals to descend among the healthy crops as well. The precision used by the LettuceBot not only promises to make usage of pesticides more efficient but, could lead to less pesticides ingested by consumers while decreasing their environmentally damaging side effects such as chemical runoff.
Regions and Innovators
At AER’s event on AI, attendees will have the opportunity learn from innovators about where their work is taking them. While it can seem like big corporations hold a monopoly on technological development, stimulating innovation in AI will allow regions to be active players in the 4th industrial revolution and create new opportunities for employment. AER’s event will provide information on how regions and innovators can get funding for research and projects, with a presentation from DG Connect’s Cecile Huet. As illustrated above, the impact of AI is cross cutting and promises to transform the way many industries are run. With a study visit to VUB’s AI lab, attendees of AER’s event will have the opportunity to witness this innovation first hand.