By Kaveesha Fernando
SILICON VALLEY, America (CWBN)_ Respected Artificial Intelligence (AI) scientist Timnit Gebru’s statement last week that Google LLC fired her following a dispute over a research paper she authored enraged the tech and academic community, who felt that her removal was indicative of mala fide practices in Silicon Valley tech companies. Well over 2500 professionals signed a letter of protest, of which more than 1000 signatures came from Google employees themselves. Attention is now being paid to her research paper to see if it sheds light on her dismissal.
The Massachusetts Institute of Technology (MIT) Technology Review examined an early draft of the paper in question to find out why it became so controversial. One of the paper’s co-authors, Emily M. Bender told the Review that the goal of the paper was to take stock of the landscape of current research in natural-language processing, a form of AI which extracts meaning from human language to inform its decisions.
The Review reports that the researchers disclose one of the main issues with large language models; it can be used to fool people. The famous case of a college student who churned out AI-generated self-help and productivity advice on a blog was cited as an example. This could lead to AI models being used to generate misinformation on topics of significance, such as elections and the COVID-19 pandemic, according to the Review.
“We are working at a scale where the people building the things can’t actually get their arms around the data,” she said. “And because the upsides are so obvious, it’s particularly important to step back and ask ourselves, what are the possible downsides? … How do we get the benefits of this while mitigating the risk?,” Bender asked.
Head of Google AI Jeff Dean is reported to have stated in an internal email that the paper ignored recent work which he claims removed some of the issues mentioned in the paper. However, his response (and Google’s move as a whole in firing Gebru) has just been seen as highlighting the issues within tech companies.
Gebru is well known in the industry for co-authoring a famous research paper which highlighted the fact that facial analysis technology worked against women and persons of colour, which could lead to discrimination. She is also known for cofounding “Black in AI”, a political action committee which focuses on increasing the presence of Black people in the field of AI. MIT Technology Review calls team she helped build at Google is one of the most diverse in AI as well as experts in their own right.
