Thanks to the convergence of big data, machine learning, and the IoT, artificial intelligence has come to the forefront of IT, insurance, financial and automotive industries. From smarter process automation to life-saving research, the potential is just beginning to be realized.
However, AI is already raising some rather disconcerting questions that we may not have the answers to just yet.
1 The Questions of Ethics
One of the most popular conundrums for AI is “the trolley problem,” which is incredibly relevant when it comes to autonomous cars. If a self-driving car must make a decision such as driving into a group of people or swerving and only hitting one person on the other side of the street, what will the car do? It is largely believed that such scenarios are incredibly unlikely, making such questions entirely hypothetical but it does raise ethical concerns about AI. If an AI device is ever faced with an ethical dilemma, how will it respond?
2 Present Biases
Ethics and biases have often gone hand in hand and while ethical dilemmas for AI may not yet be present, examples of biased AI have already been found. This is concerning as AI now helps to determine everything from legal decisions to loan approvals. For example, when using AI to select candidates to interview, candidates with European American names were 50% more likely to receive an interview for any given position. This bias can also be found in data and information. Facebook uses an algorithm, which pushes posts it believes that a user wants to see towards the top of his or her newsfeed. However, by pushing the type of posts users want, it just creates confirmation bias and echo chambers rather than giving users access to content of varying perspectives.
3 Threats to Security
Most serious are the inherent security risks of network enabled devices. Just last year, two hackers were able to remotely take over a Jeep, turning its wheels and pressing its brakes. This brings up concerns about the ability to remotely implant malicious software. What will happen when AI devices are used find vulnerabilities in critical infrastructure that no human could foresee and worse what if AI is weaponized?
Source: 2017 Beyond the Hype: Artificial Intelligence, Machine Learning and Non-Malware Attacks Report
Regardless, the benefits of AI greatly outweigh the concerns that we may have. However, this does not mean that they shouldn’t be addressed as we move forward and AI becomes further integrated into our daily lives.
To learn more about AI, contact the professionals at Ciklum today. Our experts are prepared to share with you what AI can do for your business.