With the development and integration of artificial intelligence (AI) into our daily lives, ethics and bias quickly become main concerns. Both ethics and bias also make implementing artificial intelligence in a practical sense that more complicated. Whose definition of what is ethical are corporations programming their AI systems to default to? It’s highly unlikely that there will be a standard answer to this question across companies, industries and countries; especially considering the lack of diversity in the AI design world.
The question of ethics, security and privacy is relevant enough, that IEEE (the world’s largest organization dedicated to advancing technology for the benefit of humanity) has formed an initiative to examine ethics in the design of artificial intelligence systems. The anticipated benefits of AI are incredible–for example, the economic impact of automation and life-altering advances in medical technologies. The potential consequences however, could include long-term employment consequences, implications regarding the safety of autonomous “beings”, and the measures that will need to be in place to address cybersecurity.
Artificial Intelligence Basics
A sub-field of computer science, data mining is the process of discovering patterns in large data sets involving artificial intelligence, machine learning statistics and database systems and summarizing it into useful, actionable information. This information can be used to increase revenue, cut costs, solve, diagnosis, classify or forecast problems.
The Big Players
For AI to flourish it needs to continuously harvest your personal information and preferences. Little by little AI is making its way into our homes and workplaces via browser updates, and voice activated assistants. Integrating IoT into our smart homes, smart businesses, wearable technology and into our daily business tasks or life activities it is impossible for consumers to not feed our devices –and the companies that designed them– highly personal data. Here’s how some main players in the AI field are handling that personal data:
Google lives in an AI- first world. Recently, Android’s smartphone Pixel was the first Android phone to be pre-loaded with the company’s AI assistant (Google Assistant). Coupled with Google Home it has created great buzz among consumers and technology enthusiasts. While Google’s CEO, Sundar Pichai, admits that its “always-on” assistant is still early on in its development he envisions a day where there is an individual Google, for every individual.
As the Assistant grabs intelligence from your day-to-day activities, gathers and stores your data on its servers permanently it listens to your intimate family conversations from within your home as it strives to become more powerful and faster. That same data, however, will also help Google understand who you are so that it can then advertise products that are most likely to appeal to you.
Microsoft’s personal assistant, Cortana serves you personally relevant suggestions and alerts. Microsoft will allow you to control how much data you share. Cortanta “works best when you sign in and allow her to use data from your device and connect services.” But if you don’t sign in to give it explicit permission it can still perform other tasks for you that don’t require personalization (set alarms, perform calculations or translations). Cortanta can also help with your queries as you search the web (with the edge browser).
Microsoft’s privacy agreement does clearly state that they do not use the data you share to target ads, unlike other platforms. While their privacy settings can be adjusted, their virtual assistant collects information from your device, applications and networks you are connected to by default.
As of September 2016, Microsoft has dedicated the Microsoft AI and Research Group, to making AI more accessible and valuable to everyone across apps, services and infrastructures. As Microsoft’s CEO, Satya Nadella stated, “we will be infusing AI into everything we deliver across our computing platforms and experiences”. To be determined on whether that is a good thing.
Apple has not been one to shy away from privacy concerns, and has gone head to head with government officials who they feel have crossed over the line of duty. Clearly stated in Apple’s privacy statement, “We also refuse to add a “backdoor” into any of our products because that undermines the protections we’ve built in. We’re committed to using powerful encryption because you should know the data on your device and the information you share with others is protected.”
Apple also seems to stand out from their competitors in that the information used to provide suggestions is kept on your device and not in the cloud. Siri can only associate your HomeKit devices with your anonymous Siri identifier, not you personally. Any applications developed for HomeKit are restricted by Apple’s guidelines to using data solely for home configuration or automation services. Data related to your home is stored encrypted in the keychain of your device. It’s also encrypted in transit between your Apple device and the other services or IoT accessories you’re controlling.
Things to Consider
One must consider the social, technological, legal, and philosophical questions surrounding artificial intelligence. The best ways to manage the associated risks and rewards that result from AI in the practical world will have to be fully developed. As technological development of AI grows, what kinds of guiding principles, regulations or even preemptive bans should be considered across industries and international borders?
We will soon see how AI adapts to unforeseen situations, as it continues to transition from a laboratory environment into the outside real world where unpredictable things can happen.Exploring the future possibilities of what artificial intelligence can offer us, is an exciting endeavor. But with the building of personalized, customizable assistants and mastering of AI comes at the price of little to no privacy. The real benefits to be received by the average user’s exchange of personal data to these companies is yet to be determined.
(image via bluediamondgallery)