As our technological landscape changes shape in the 21st century, so do the rules that go with it. Intellectual property laws have expanded to encompass online material, data protection measures have been introduced to help users avoid scams and scandals, and entire industries have been built on preventing cybercrimes.
And when it comes to artificial intelligence – that’s a whole can of worms in itself.
These days, AI is almost everywhere. To borrow from a report by The Verge at the start of this year, “It’s being deployed in health care and warfare; it’s helping people make music and books; it’s scrutinizing your resume, judging your creditworthiness, and tweaking the photos you take on your phone. In short, it’s making decisions that affect your life whether you like it or not.”
AI technology is ubiquitous, omnipresent, and, unfortunately, a little ethically ambiguous.
This was proven quite succinctly in 2016, when Microsoft’s well-meaning AI chatbot was taught to spew racist responses within just 24 hours of being launched on Twitter. It didn’t know what it was doing was wrong, of course – it was just going along with what its machine-learning algorithm had been programmed to do.
What’s more, research has found that AI is capable of showing inherent bias towards or against certain groups of people. According to The Independent, “researchers found male names were more closely associated with career-related terms than female ones, which were more closely associated with words related to the family,” and “there were strong associations, known as word ‘embeddings’, between European or American names and pleasant terms, and African-American names and unpleasant terms.”
The problem is not that computer systems are bias, but that we – the people teaching them – are. So, how do we go about ensuring AI is as ethical as possible? SAP has a number of guiding principles:
Stick to core values and maintain transparency
SAP has a dedicated AI Ethics Steering Committee, which is in place to ensure that any and all digital technology is engineered to comply with the company’s core values. “Where there is a conflict with our principles, we will endeavour to prevent the inappropriate use of our technology,” the company says.
These values are available for anyone to read about in SAP’s code of business conduct, which has made public to ensure they are held to accountability.
Design with people in mind
“We strive to create AI software systems that are inclusive and that seek to empower and augment the talents of our diverse usership,” SAP explains. For this reason, all AI systems are designed and developed “in a collaborative, multidisciplinary, and demographically diverse environment.” This also helps prevent bias, as the technical teams building the AI take into account the information they are using as a foundation for the system and make a conscious effort to avoid negative associations with particular genders/races/orientations etc.
Have high quality and safety standards
As with any other product or application, AI ought to undergo rigorous trials and testing before being rolled out to customers. “Our AI software undergoes thorough testing under real-world scenarios to firmly validate they are fit for purpose and that the product specifications are met,” SAP says. If at any point the AI misbehaves during testing – or, indeed, after it has been rolled out – SAP is prepared to rework the issue until it is resolved.
These standards also apply to data protection, which SAP has identified as another potential pain point for AI. For this reason, they “communicate clearly how, why, where, and when customer and anonymized user data is used in our AI software.”
Acknowledge the challenges
When you’re already at the forefront of business technology, it’s easy to become hubristic about other potential challenges. However, SAP has made a conscious effort to “engage with the wider societal challenges of AI”, and accept that it is still something that needs to be approached with caution.
Before we know it, just about every business in the world will be utilising AI in some form or another. We won’t get it right straight away, and we certainly won’t make progress without some teething problems, first.
But while we’re still learning, it’s imperative that we do our best to practise good digital ethics – and SAP’s guidelines are a fantastic place to start with that.