Google’s AI journey has hit a major hurdle recently as its AI image generator received a lot of backlash and accused of being racist with content generated with twisted historical facts. The company was understandably forced to stop its public release as it looks to fix the issues that have brought these negative marks.
But now Google CEO, Sundar Pichai has finally spoken about the matter in front of his employees and reportedly told them these are ‘unacceptable errors’. Pichai has spoken about the issue in an internal memo, as quoted by a report this week, saying the company will make a whole list of changes and improvements in the AI mechanism to ensure these issues and mistakes don’t happen ever again.
“We’ll be driving a clear set of actions, including structural changes, updated product guidelines, improved launch processes, robust evals and red-teaming, and technical recommendations. We are looking across all of this and will make the necessary changes,” the memo had Pichai saying. The Google chief did admit that no AI is perfect but the world expects a giant like Google to not make these mistakes, setting the bar much higher for the company.
He claims the changes made to the AI tool are already showing improved results for a wide range of prompts but we’ll probably have to wait for a few weeks before the new-look AI generation tool is rolled out once again for the public to see if these changes are actually effective and fix the errors made by the AI tool earlier.
Google has repeatedly talked about going slow with the AI roll out and instances like these are a clear sign that these tools need more learning and data training, which includes understanding the context before dishing out the results for these prompts. Pichai is under a lot of pressure to get things right, or faces the prospect of losing his position as the head of Google in the near future.
Comments
0 comment