By Mike Robuck
Kaspersky highlights ethics as key for AI success
The promise of AI in a 5G world seems unlimited, but there needs to be firm guidelines in place in order for it to reach its full potential, according to an industry expert.
Kaspersky Senior Data Scientist Vladislav Tushkanov told Mobile World Daily that AI ethics need to be in place across academic ethicists, researchers, operators, lawmakers and activists, but the final results are always a compromise.
AI ethics comes with specific issues that could hinder ethical inquiry into the algorithms by academic researchers. As an industry, operators and the rest of the AI ecosystem have to understand how to measure the impact of issues in the field and on humans, according to Tushkanov.
“Just like industrialisation has led to a climate change crisis, irresponsible adoption of AI can lead to sad consequences for society, which is why there is a need to address it in the form of standards and ethical guidelines,” Tushkanov stated.
There are three main challenges that operators face across their own AI-based operations and consumer facing innovations. The first problem is exaggerated expectations for AI. For example, when a supposedly smart AI customer support chatbot drives down customer satisfaction “disillusionment follows” for the users.
The second challenge is that machine learning, which is the forebearer of AI, can be brittle and fail in unexpected ways. The third challenge is centred on the quality of data. Having the wrong assumptions can lead to bad performance as well as biased algorithms.
On the plus side, with greater 5G connectivity, there are opportunities for better analytics and enhanced decision-making across logistics and transportation, urban planning and industry.
“If machine learning technologies are deployed responsibly, securely and in a human-centered way, there can be great synergy,” added Tushkanov.