Microsoft to stop selling AI facial recognition software
- The decision was outlined in the Microsoft Responsible AI Standard
- Amazon and Microsoft have both eases sales of the software to US police agencies
- Small companies continue to sell AI-based facial softwares
The company that created the Windows Operating Software has decided to stop selling their artificial intelligence-based facial-analysis tools after finding that the algorithms were prone to problematic biases and inaccuracies, Bloomberg reported.
The tools which were capable of identifying an individual’s gender, mood and even emotional state are going to continue to be available to existing users for a year before they expire. For access to the tech giant’s Microsoft Azure Face API and Computer Vision and Video Indexer, new customers will need to apply for access. Details of the decision, along with the changes to access were made available in a blog outlining the second update to Microsoft’s Responsible AI Standard written by the Chief Responsible AI Officer Natasha Cramption and the Azure AI Product Manager Sarah Bird.
That isn’t to say that Microsoft is completely doing away with using AI to help gauge human reactions. According to Bloomberg, the Satya Nadella-led company was building an AI for sales representatives to run analysis on how customers engage in teleconferences. The software is meant to analyse what way potential customers might be reacting to a pitch.
The change comes nearly two years after both Microsoft and Amazon ceased sales of their facial recognition and AI services to American police forces after research indicated that the software performed badly on individuals with darker skin. Some states across the country have pushed out laws dictating the use of such technology. Washington, where both companies are headquartered, is one such state. California and Vermont have banned the software outright whereas other states like Massachusetts and Oregon have banned them in select cities.
While big tech companies seem to be backing away from such technology in fear of facing backlash, smaller organisations have capitalised on these fears. Companies like NEC Corp. and Clearview AI continue to offer up facial-recognition tools, often courting controversy and raising questions about the invasion of privacy and oversight.