Our commitment to creating ethical AI
As artificial intelligence permeates further into every aspect of daily life, there is growing concern that humanity is becoming increasingly ‘technologized’. Only last week, officials from the Vatican, European Union, and United Nations, along with IBM and Microsoft signed the “Rome Call for AI Ethics”, a joint statement calling for the development of ethical guidelines for AI.
The question of what role AI should play in society has real and significant consequences for the future. On the one hand, it’s undeniable that AI has the potential to solve many of the most pressing issues faced by humanity today. For example, the Director of the UN Food and Agriculture Organization (FAO) has said that “AI has the transformative power to make food systems more efficient, sustainable and inclusive.”
But the pendulum can easily swing the other way. The recent introduction of facial recognition technology demonstrates this clearly, as campaigners fear AI could be used in these circumstances to curb freedoms and erase the right to privacy of citizens.
The role of ethics in AI also touches upon our work at Anyline. Some of our own solutions can be used to scan IDs and passports, which is used by many different industries, including transport, law enforcement, and the banking sector, to name a few.
So how do we ensure our own ethical standards?
Firstly, we ensure that all data collected is processed entirely on device. This protects the data from interception and ensures privacy. Secondly, we ensure our products comply with the General Data Protection Regulations (GDPR), and collect no personal information from any ID document scanned with our solutions. Finally, we challenge ourselves to ensure that each of our products align with the ethical standards of our team.
As we develop more innovative solutions, we in the computer vision community would be wise for us to heed the advice of the Rome Call of AI Ethics, and ensure that we are working to humanize our AI, rather than the other way around!
