For firms such as Microsoft and Google, AI is a big part of their future, providing methods to improve current products and make whole new income streams. But, as disclosed by latest fiscal filings, both companies also claimed that AI (specifically biased AI that makes bad moves) can possibly harm their businesses and brands.
These filings were made in the firms’ 10-K forms. These are standardized papers that companies are lawfully needed to file each year, offering investors a broad insight of their recent finances and business. In the section named “risk factors,” both Alphabet and Microsoft mentioned AI for the first time.
These filings are not, in general, hugely shocking. The idea of the “risk factors” section is to keep investors well-versed, but also alleviate future court cases that may blame management of hiding possible issues. Due to this they aim to be very broad in their work, offering even the most understandable manners a business can go wrong. This may comprise issues such as “somebody made an enhanced product as compared to us and now we do not have any users,” and “we invest all our money so now do not have any left.”
Speaking of AI, the newest AI software by DeepMind, AlphaFold, earlier won the CASP (Critical Assessment of Structure Prediction) contest by precisely forecasting the 3D structures into which proteins can be molded. AlphaFold was declared as the winner amongst a sum of 98 algorithms. Out of 43 proteins, it forecasted the shapes of 25. In comparison, the first runner-up algorithm could forecast just 3 of the 43 structures of protein, as per the media.
The competition was organized by the Protein Structure Prediction Centre, an agency backed by the National Institute of General Medical Sciences in the U.S. “This is a really a major moment for us,” claimed CEO and co-founder at DeepMind, Demis Hassabis, to the media in an interview.